ml-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "84c08e88",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:03:58.618871Z",
     "start_time": "2021-05-03T14:03:57.746396Z"
    }
   },
   "outputs": [],
   "source": [
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "Created on Sat Apr 24 12:43:53 2021\n",
    "\n",
    "@author: karan\n",
    "\"\"\"\n",
    "\n",
    "\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from matplotlib import pyplot as plt\n",
    "# importing data\n",
    "df = pd.read_csv(\"Training Data.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30cddfb1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-04-24T15:07:37.573869Z",
     "start_time": "2021-04-24T15:07:37.556938Z"
    }
   },
   "source": [
    "## Data preprocessing\n",
    "- Deleting Coloums\n",
    "- Removing Rows Contaning Null values\n",
    "- Then Dividing Dataframe into <b> X and Y </b>\n",
    "- Then Dataset into **Test** and **Train**\n",
    "- Then **Feature Scaling**\n",
    "\n",
    "### Then Finally We have\n",
    "* <b> X_Trian , X_Test , Y_Train , Y_Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "479f51a3",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:03:59.556903Z",
     "start_time": "2021-05-03T14:03:58.619832Z"
    }
   },
   "outputs": [],
   "source": [
    "# Colum del\n",
    "df.drop([\n",
    "    \"clmbuserid_cr\", \"conversiontime_cr\", \"imprid_cr\", \"audiences_cr\",\n",
    "    \"clickbid_cr\", \"ip_cr\", \"refurl_cr\", \"templateid_cr\", \"templateid_cr\",\n",
    "    \"time_cr\", \"v_cr\", \"allAudiences_cr\", \"impressionTimeInMillis_cr\",\n",
    "    \"stateGrpDimId_cr\", \"deviceId_cr\", \"uvh_cr\", \"uv_cr\", \"platformId_cr\",\n",
    "    \"sdkVersion_cr\", \"usrClusterId_cr\", \"cityGrpDimId_cr\", \"siteClusterIds_cr\",\n",
    "    \"refClusterId_cr\", \"paid_cr\", \"spend_cr\", \"attributionType_cr\",\n",
    "    \"conversionid_cr\", \"optimize_on_cr\", \"bundleId_cr\", \"clickTimeInMillis_cr\"\n",
    "    \n",
    "],\n",
    "        axis=1,\n",
    "        inplace=True)\n",
    "\n",
    "# Not REmoving\n",
    "# \"connTypeDimId_cr\"\n",
    "\n",
    "# Convert Values to INT in Y\n",
    "df['conversion_fraud'] = df['conversion_fraud'].astype(int)\n",
    "# Droping Null value\n",
    "df = df.dropna(how='any', axis=0)\n",
    "\n",
    "# Dividing Dataset INTO X and Y\n",
    "X = df.iloc[:, :-1].values\n",
    "y = df.iloc[:, -1].values\n",
    "\n",
    "# Splitting the dataset into the Training set and Test set\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25,random_state=0)\n",
    "\n",
    "# Feature Scaling\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "sc = StandardScaler()\n",
    "# sc = MinMaxScaler()\n",
    "X_train = sc.fit_transform(X_train)\n",
    "X_test = sc.transform(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edaf2496",
   "metadata": {},
   "source": [
    "# Logestic Regresion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "583d5206",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:03:59.634770Z",
     "start_time": "2021-05-03T14:03:59.557901Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[209   2]\n",
      " [ 18   7]]\n",
      "Logestic Regresion Accuracy :  0.9152542372881356\n"
     ]
    }
   ],
   "source": [
    "# Fitting Logistic Regression to the Training set\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "LR_classifier = LogisticRegression()\n",
    "LR_classifier.fit(X_train, y_train)\n",
    "\n",
    "############################### Evaluating the model\n",
    "# Predicting the Test set results\n",
    "y_pred = LR_classifier.predict(X_test)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "\n",
    "cm1 = confusion_matrix(y_test, y_pred)\n",
    "print(cm1)\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "ac1 = accuracy_score(y_test, y_pred)\n",
    "print(\"Logestic Regresion Accuracy : \", ac1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "403c9f61",
   "metadata": {},
   "source": [
    "## K-NN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d4454393",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:00.291736Z",
     "start_time": "2021-05-03T14:03:59.636764Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'correct True and False')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "################ K-NN\n",
    "# Fitting K-NN to the Training set\n",
    "list_true = []\n",
    "list_flase = []\n",
    "list_i = []\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "for i in range(1, 30):\n",
    "    knn_classifier_test = KNeighborsClassifier(n_neighbors=i,\n",
    "                                               metric='minkowski',\n",
    "                                               p=2)\n",
    "    knn_classifier_test.fit(X_train, y_train)\n",
    "\n",
    "    y_pred = knn_classifier_test.predict(X_test)\n",
    "\n",
    "    # Making the Confusion Matrix\n",
    "    from sklearn.metrics import confusion_matrix\n",
    "    cm2 = confusion_matrix(y_test, y_pred)\n",
    "    from sklearn.metrics import accuracy_score\n",
    "    list_i.append(i)\n",
    "    list_true.append(cm2[0][0])\n",
    "    list_flase.append(cm2[1][1]*10)\n",
    "plt.plot(list_i, list_true, \"g\")\n",
    "plt.plot(list_i, list_flase, \"b\")\n",
    "plt.xlabel(\"value of i\")\n",
    "plt.ylabel(\"correct True and False\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af21b761",
   "metadata": {},
   "source": [
    "## 17 is best n_neighbors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "06c40178",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:00.323650Z",
     "start_time": "2021-05-03T14:04:00.292734Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K-nn :  0.9110169491525424\n",
      "[[201  10]\n",
      " [ 11  14]]\n"
     ]
    }
   ],
   "source": [
    "knn_classifier = KNeighborsClassifier(n_neighbors=17, metric='minkowski', p=2)\n",
    "knn_classifier.fit(X_train, y_train)\n",
    "\n",
    "y_pred = knn_classifier.predict(X_test)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "cm2 = confusion_matrix(y_test, y_pred)\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "ac2 = accuracy_score(y_test, y_pred)\n",
    "print(\"K-nn : \", ac2)\n",
    "print(cm2)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5fd3421",
   "metadata": {},
   "source": [
    "## SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "af241da7",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:00.385490Z",
     "start_time": "2021-05-03T14:04:00.324648Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "svm rbf :  0.9194915254237288\n",
      "[[211   0]\n",
      " [ 19   6]]\n"
     ]
    }
   ],
   "source": [
    "############ SVM rbf\n",
    "# Fitting SVM to the Training set\n",
    "from sklearn.svm import SVC\n",
    "\n",
    "svm_rbf_classifier = SVC(kernel='rbf', random_state=0)\n",
    "svm_rbf_classifier.fit(X_train, y_train)\n",
    "\n",
    "y_pred = svm_rbf_classifier.predict(X_test)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "cm3 = confusion_matrix(y_test, y_pred)\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "ac3 = accuracy_score(y_test, y_pred)\n",
    "print(\"svm rbf : \", ac3)\n",
    "print(cm3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7d5591ab",
   "metadata": {},
   "source": [
    "## Desision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "19cc2ded",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:00.432181Z",
     "start_time": "2021-05-03T14:04:00.388481Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Desision Tree:  0.9067796610169492\n",
      "[[199  12]\n",
      " [ 10  15]] \n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Fitting Decision Tree Classification to the Training set\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "dt_classifier = DecisionTreeClassifier(criterion='entropy')\n",
    "dt_classifier.fit(X_train, y_train)\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred = dt_classifier.predict(X_test)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "cm4 = confusion_matrix(y_test, y_pred)\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "ac4 = accuracy_score(y_test, y_pred)\n",
    "print(\"Desision Tree: \", ac4)\n",
    "print(cm4, \"\\n\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6f3aab9",
   "metadata": {},
   "source": [
    "## Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "e5f9f1a5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:01.213902Z",
     "start_time": "2021-05-03T14:04:00.434176Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'correct True and False')"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnEAAAGpCAYAAADiCGDnAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAABVaUlEQVR4nO3dd5iU5fX/8fdZehOkCkqzF1RQVBR7B42osWBBYPVr8tMkGo01xphiYmIsMdXCAiqosRtlY8GGXQQUERVUVGQWEBBY+u6e3x/3rFnJltndeeaZ2fm8rmuunXnmKWd3dnbP3OXc5u6IiIiISG4piDsAEREREak/JXEiIiIiOUhJnIiIiEgOUhInIiIikoOUxImIiIjkoOZxB5AJXbt29X79+sUdhoiIiEid3nnnna/dvVtd++VFEtevXz+mT58edxgiIiIidTKzz1PZT92pIiIiIjlISZyIiIhIDlISJyIiIpKDlMSJiIiI5CAlcSIiIiI5SEmciIiISA5SEiciIiKSg5TEiYiIiOQgJXEiIiIiOUhJnIiIiEgOUhInIiIikoOUxImIiIjkICVxIiIiIjlISZyIiIhIDlISJyIsXw5lZXFHkbqVK2HDhrijEJFstWpV3BFkhpI4kTy3ciVstx385jdxR5KasjIYOBDOPDPuSEQkGxUXQ9eu8K9/xR1J9JTEieS5+++Hb76BO+/Mjda4Z56BBQvgkUdg6tS4oxGRbLJxI1x8MWzaBJdeCmvXxh1RtJTEieS5oiJo0wYSCXj66bijqVtREXTrBv37w09/mhuJp4hkxt/+Bh9/DFddBQsXwo03xh1RtJTEieSx99+Ht96C666D7t1DgpTNli6FJ56AUaPCH+fZs+Guu+KOSkSywdKl8KtfwTHHwPXXw2mnwR/+AF9+GXdk0VESJ5LHxo+H5s1h7NiQGD3xRPhDmK0mTQrdJGPHwsknwyGHwC9+EbqDRSS/XXstlJbCzTeDGfzxj+AOV14Zd2TRURInkqc2boR77oETTgjdk2PHhq7Je++NO7LquYeWwn33hQEDwh/pW2+FZcvg17+OOzoRidN778Edd8AFF8Cuu4ZtffvCz34GkyfD66/HG19UlMSJ5KmnngqtboWF4fFuu8F++8G4cSFhyjbvvBO6TyvjhTBL9bzz4C9/gY8+ii00EYmRexgf26lTGBpS1RVXQK9ecNFFUFERR3TRUhInkqeKiqBnzzB+pFJhIcyZA9OnxxdXTYqKoHVrGDnyu9t/+1to2zbMRBOR/PP44/D882E8XOfO332ufXu44QZ4++3s7WVoDCVxInlo0SKYMgVGjw5j4iqdfnqYqZptExzWrQtdIqecAh07fve57t3DuLinnsqN2bUikj4bNoQu0113hR/+sPp9zjorDMO48sowZq4pURInkofuuSd0LYwd+93tHTuGRGny5Oyqr/Too6EocdWu1Kp+8hPYfvvQpbJpU2ZjE5H43HYbfPIJ3HLLdz+QVlVQAH/+cyijdMMNmY0vakriRPKMe5iVeuCBsOOO//t8YWFYsubRRzMfW02KikJduEMOqf75li3hpptg7lz45z8zG5uIxGPx4rDSzPHHw9FH177vkCGhRe5PfwrFwpsKJXEieeb118MkgJpatQ4+OCRM2dKlumBBWJlhzJjwibom3/seHHkk/PKXYS1YEWnarrkmDLW46abU9v/978PfkMsvjzauTFISJ5JnioqgXTs49dTqny8oCN2szz+fHZ9YJ04M5URGj659P7NQH2rlyv+doSYiTcvMmWEm/Y9/XH2PQnV69w6zVR98EKZNiza+TFESJ5JHSkvhgQfCBIb27Wveb/TokBRNmJCx0KpVURG6fo88MtR8qsvuu8MPfgB//zt88EH08YlI5rmH9VG7dAkFfuvjsstCMnfxxU2j5IiSOJE88tBDIZHbfELD5vr0gaOOCglUnH/oXngBPv+85q7f6vz619ChA1xySXbWuxORxnn4YXj55TAerlOn+h3btm1YimvGjPg/pKaDkjiRPFJUBDvsAEOH1r1vYSF88UXoVo1LUVH4I33iiakf07VrGBf39NOhjIqINB3r14fWtN13D4W+G2LkSDjgALj66jCJK5cpiRPJEx9/HMaBFBaGrtK6jBgBW24Z3wSHFSvCJ+6zzgpFfuvjwgthp51Ca9zGjdHEJyKZd/PNYazurbfWXFKkLpVL9i1eDL/7XRqDi0FkSZyZ9TazF8zsAzObY2YXJbefmnxcYWaDq+zfz8zWmdms5K3aQgFm1tnMnjWzecmvW0b1PYg0JRMmhEkL55yT2v6tW4cE6pFHQkKVafffHwp51qcrtVKLFuGP/ccfw9/+lv7YRCTzEomQdJ14Ihx+eOPOtc8+YezvLbeEOnO5KsqWuDLgUnffFRgCXGhmuwLvAycDL1dzzCfuPjB5q6H2MlcCU919B2Bq8nHs1q2LOwKRmpWVhVmew4aFdQRTVVgYEqn77osutpoUFcGee8KgQQ07fvhwOPbYsBTP0qXpjU1EMu/qq0Mx7z/9KT3n+93vwge+yy5Lz/niEFkS5+4Jd5+RvL8amAts7e5z3b0xS1WPACYm708ETmxUoGnwl79At26wZk3ckUg2WLQI9t47LAOVLZ55JsRV31atQYPCIvOZ7lJ9772wfmuqXb81ufnmMJGjvjPY0mnZspBMjhiR++NvpPHcw4D8XXYJ63lKaqZPD70JF18M222XnnP26hUSw0cfDZOoclFGxsSZWT9gEPBmHbv2N7OZZvaSmR1Uwz493D2RvF8C9Kjhmueb2XQzm7404o/hu+wSErhc/SWQ9Lr66jDz6YILsqeFtqgoDPg//vj6H1tYCO+8A+++m/64ajJ+fFiF4ayzGneeXXYJ4+PuuCMkhpn26adhAPULL4RJFgcdBAsXZj4OyQ4bN4aZ4ddeC19+CYceCk88EXdU2a+ypEiPHvDzn6f33JdcEsoXXXwxlJen99yZEHkSZ2btgYeBi929ts+hCaCPuw8CLgEmm9kWtZ3b3R2otoiAu9/h7oPdfXC3bt0aGH1qDjooFE8tLo70MpID3n47dFsefXSY2ZmuZv/G+Prr8I9i1KiQGNXXmWeG48aPT39s1dm4MaztOmJEqAPVWL/8ZZjhmumSI2+9FZb6+frrsOLElCnw2WdhWyYTYskOK1eGLv6JE0MX//z5YdH2k07SuM26PPAAvPoqXH89bFFrVlB/rVvDjTeGD3l33ZXec2eEu0d2A1oATwOXVPPci8DgWo6t9nngI6Bn8n5P4KO64th77709aiec4N6vn3tFReSXkixVUeF+wAHuPXq4r1rl/v3vu7dt675wYbxx3XqrO7jPnt3wc5x2mnuXLu7r16cvrpo89FCIt7g4fef8y1/COR97LH3nrM2jj7q3aeO+7bbuH3743+3vvee+zTbuHTq4/+c/mYlF4vf55+677ebevLn7xIn/3V5aGv53gPull7qXl8cXY7Zas8a9d2/3gQPdy8qiuUZFhftBB7l37eq+YkU016gvYLqnkmelslNDboABdwO31vD8d5I0oBvQLHl/W+AroHM1x90IXJm8fyXwx7piyUQS949/hJ/m3LmRX0qy1OTJ4Xdg3Ljw+NNP3Vu1cj/77Phiqqhw33139332adx5iovD9/bgg+mJqzbDh7tvvXV6/2Bv2uS+667u220XfSJ6663uZu777ee+ePH/Pr9wofuee7o3a+Z+113RxiLxmzHDvWdP944d3adO/d/ny8rcf/zj8P465RT3tWszHmJW+9Wvws/mpZeivc6MGeF9e+ml0V4nVdmQxB1I6Op8D5iVvA0HTgIWAhuAxcDTyf2/D8xJ7jcD+F6Vc91VmfABXQizUucBz1WX6G1+y0QSt2BB+GnefHPkl5IstGZNaGHZa6/vfpq+6qrwe/H66/HENX16uP4//tG485SVhcRq2LD0xFWThQvdCwrcf/7z9J/76afDz+KPf0z/ud3Dz+iii8I1Tjop/E7UZNUq92OPDfv+/OdqwW+qnnrKvV079z593N9/v+b9KirC/w4z9/33d1+yJHMxZrMvvwy9GaeckpnrnXuue4sW7h9/nJnr1Sb2JC6bbplI4tzDJ/0jj8zIpSTLXHddeDe9/PJ3t69a5b7VVqFVJo6ukgsucG/dOj1dBD//eUiwvvyy8eeqye9+F36O8+ZFc/7jjw9dmSUl6T3vmjUhcQP3iy9OrRVx40b3884Lx5x5Zma6qiVz/vnP0No6aJD7okWpHfPQQ+H9uv322ZFIxO2ss0JvxmefZeZ6JSXh78P3vpeZ69VGSVwMSdyll7q3bOm+enVGLidZ4osvwvin006r/vnx48M77Z57MhqWr13r3qlT+EOYDvPmhe/j+uvTc77NVVSEf16HHBLN+d3dP/oofNI+77z0nXPx4pCkm4Wu1PqoqAg/Twjf9/Ll6YtL4lFe7n7lleE1HT68/v8PXnstjM3q0sX91VejiTEXvP56+BlefXVmr/uHP4TrPvNMZq+7OSVxMSRxU6eGn+jjj2fkcpIlzjwzfFpcsKD658vL3QcPDt2RpaWZi6tyjF5143Aa6pBDQqIVRffftGkh3gkT0n/uqi65JCRcM2Y0/lwffRQmL7RpEyYzNNSkSeED4C67ZK7VQdJv/Xr3kSPD7/EPfhDGYjbEvHnhfdaqVWbGoWab8vLwwahnz8w3iqxfH8bO7rZbw1+/dFASF0MSt2GDe/v27j/8YUYuJ1ngtdf823FNtXnllbDfL36RmbjcQ9d+v37p7cadONGr7TZOh7Fjw/sn6kR3xYrQ0nHwwY1LRqdNc+/c2b1bN/c33mh8XC++GFpOe/Rwf/vtxp9PMmvZsjDDEUJrTmM/6CxdGma7m7n/6U/5NW7ynnsy84GuJo8+Gq7/17/Gc313JXGxJHHu7ieeGAax5tMbLl+Vl7vvu2/qnxZHjgzjXT7/PPrYPvss/PG/7rr0nre0NIwZGTMmveddtSoMAE9nN2dt/vlPb9Rs2/vvD60kO+3k/skn6Yvrgw9C4t22rfsTT6TvvBKtTz4JvwstW4bfjXRZu9b91FPD7+qFF0ZXYiOblJaGXot99omv5EpFhfvhh4cPacuWxRODkriYkrjbbw8/1TlzMnZJicndd4fXumrdp9p8/nlI4kaOjDYu95C8mdXcxdsY//d/IclYtSp95xw3LvwsX3stfeesTVmZ+x57hIRp3brUj6uocL/hhhDrgQdG8we+pCR0vxcUxNsSIKl5883QGrvlltG0UJeXu//sZ+F37nvfy+yQjDj84hfhe417POC774b34E9+Es/1lcTFlMR98UX4qd54Y8YuKTFYvdq9V6/QElefT4uVf6BeeSW62MrL3fv2jW6mdOWA43TWOBs61H3nnTPbgv388+H7+N3vUtt/06YwVALcTz+9fslffZWWhn/YKgKb3SqLOvfv/92izlH4299CUjF4sHsiEe214rJgQfige8YZcUcS/OAHYYbxBx9k/tpK4mJK4tzdBwwITbHSdF1zjTeo5aiyq2DvvaP7x/zccyG2yZOjOX9FRRiAf8AB6Tnfhx96pPXbanPiiaEbt64SEKtXux93XIjziisyk1SVlbn/6EeuIrBZ6s9/Dq3d++5bfVHnKDzxRGgF79s3nsQiaqefHpLiTAw5ScWSJaFI87HHZv7aSuJiTOIuvzyUMUhnd5Nkj8pPi2ee2bDjKwftjh+f1rC+deaZYYB8lP/0b7zR07ZCyRVXhE+7cbQuzJ8fxjHVNsZv0aJQxLmgIIyly6SKCvebbgrJwgEHhMHuEq+yslALEMKHgNqKOkfh7bfD5JdOncJkmKaicnb6tdfGHcl33XRTiGvKlMxeV0lcjEncCy+En2xjSg5I9jrttPBp8YsvGnZ85fT5rbZKf6K/YkVIMC+4IL3n3VxJSUi8Lr+8cefZtCn8HE44IT1xNcTll4f3a3UzQt9/P0xUatcuVN+PS9UisFEVQpa6VS3qfNFF8U00+Oyz0BreooX7vffGE0M6lZeH3olMl2FKxYYN7jvuGIZ7bNyYuesqiYsxidu4MczgO//8jF5WMuDll8O7prGzPt94I5znqqvSE1elyjV8p09P73mrc8IJoUWgMX/Y/v1vz+jC9NVZuTJ8H0OHfndM3tSpoStlq63c33kntvC+9dproQBsvheBjUtjijpHYfly90MPDe+f3/42tysiVBZEz9aEtPLvVCZfdyVxMSZx7u4nnxzW0szlN5Z8V3l56FbbZpv0dKGcfXYoU/Hpp40/V6V99gmzLjPxe/fYY+EvSGNKYZx0knv37pn9hFudu+4K38t994XHd98dWjl22y17xue4h6WY8rkIbFwqizq3bu3+yCNxR/Nf69eHFVkglOeJ+33UEJVLEw4Zkr3/Lysq3I8+OnRhZ2pIg5K4mJO4O+8MP93ZszN+aYlIZRmMSZPSc76FC8Mg5e9/Pz3ne+89z+inxY0bQwJ24okNO37xYvfmzUP5hLiVlYU1Lnv3/u8M4sMPT8+as+m2dGlYJD0fi8DGId1FndOtouK/E62OOSa0LOeSq64Ksb/5ZtyR1G7OnDCEJOqhKpVSTeIs7Nu0DR482KdPn57Ra371FWyzDfzhD3D55Rm9dJ2++AKuuw6uuQa23TbuaHLDqlWw447h5/Xqq2CWnvP+5jdw7bXw4otwyCGNO9cll8Bf/wqLFkHXrmkJr06XXQa33goLF0KPHvU79uab4dJLYc4c2HXXSMKrl2nT4OCDw/1zzoE774SWLeONqSbr1oUYH3oI9tsP2rWLO6KmyR1eew369oUpU2C77eKOqGbjxsEPfgC9e+fW3/VXXoHTT4e77447krr9+Mfw97/Du+/CgAHRXsvM3nH3wXXupyQuOnvuCZ07wwsvZPzStbroIrjtNujWDZ58EvbdN+6Ist+VV4aE/K23YJ990nfetWth552hSxeYPh2aNWvYeTZuhK23hkMPhQcfTF98dfngA9htN/jTn0JClir38Edwiy3g9deji6++fv1raNs2fC/pStSjUlEBv/0tPPts3JE0bX36hL+XXbrEHUndnn0Wbrgh/D3IFV27hsSoZ8+4I6nbsmVw+OFwyy3ha5SUxFURVxJ31VXhn9uyZeGfVTbYsAF69YI99ggtcokETJ4MJ54Yd2TZ69NPYZddYORImDgx/ee//34444zQ8nPeeQ07xyOPwPe/H1oLhg1Lb3x12X//0FL5/vupJz5vvRVakO64A/7v/6KNT0QkXdwz8wEv1SSuIPpQ8tewYVBWBs89F3ck//XEE7B8eWhZev31kMydfDL8+c9xR5a9LrsMWrSA3/8+mvOffjoMHQo//3lIhhqiqCi0xB19dHpjS0VhYWiRe+ut1I8pKoI2bcL3LiKSK7KthV5JXIT23x86doTi4rgj+a+iojBW78gjoXt3eP55GDECLr443MrL444wu7z4Ymjluuqq0IIZBbMwrmzJktA9Vl+LFoXfsdGjG94d2xinnx4SsvHjU9t/7Vq47z445ZTsaaEWEclFSuIi1KIFHHVU+AebDb3WX34JTz8NY8b8959927ZhcPRFF4XWuFNPDf9kJSS0F18cBjVfckm01xo8OLwut94K8+fX79i77w7jo8aMiSCwFGyxRfi9ue++1H53Hn00tDgWFkYfm4hIU6YkLmLDhoWZqrNnxx1J+Gfv/r//7Js1C8nDrbfCY4+FAZtLlmQ+vmwzblyYhfTHP4aWpqj97nfQqhX87GepH+MeWlcPOgh22CG62OpSWBgSs0ceqXvfoqIwe65yJqiIiDSMkriIHXts+DplSrxxVFSEf56HHlrzNPmLLgr/hN97L3QFf/xxRkPMKitXhhIsBx0UWpkyoWdPuPpqePxxmDo1tWNefRXmzYu/Vevgg8PvVVFR7ft99lnowh87Fgr010dEpFH0ZzRivXrBwIHxj4ubNi3Msqzrn/2JJ4aSKKtXh0TulVcyEl7W+c1v4OuvQ+tkJgey/vSn0L9/6MYtK6t7/6IiaN8+jC+Lk1lo4X3hhfB7VpMJE8K+o0dnKjIRkaZLSVwGDB8eWky++Sa+GIqKoEOHUIaiLvvtF2audu0aJkA88ED08WWTefNCXaixY2GvvTJ77dat4cYbQ7mOO++sfd/SUvjXv8LEgvbtMxNfbUaPDgnahAnVP19eHiY/HH10KEgqIiKNoyQuA4YNC//A4io1smpVKAB7xhlhIkMqttsuVCrfZ59QH+0Pf8iOyRmZcOmlIZm6/vp4rn/yyWH1hl/8ovbE/8EHYc2a+LtSK/XuHRK0CROqn+X8/PNhck22xCsikuuUxGXAkCHQqVN84+IeeCAs01Pff55duoQK4KefHurKXXBBal18uezZZ+Hf/w4127baKp4YzEJF8OXLwwoCNSkqgp12Ct3e2aKwMCRq1Y3pKyoKK5iMGJH5uEREmiIlcRnQvHloofjPf+JpzSoqCmtTNmR5rdatw4oOV1wB//xnGDNXWpr2ELNCWVkYk7bttmFMWpwGDYJzz4W//AU++uh/n//44zBesbAwu4pPjhgRErXNJzgsXx5Ki5x1VpiBKyIijackLkOGDw9LXL37bmav+8EH8MYbjftnX1AQ1uP75z/DBI1DDgnfS1Nzxx1hMfYbb8yOROO3vw2lTapbk3T8+FAaZtSozMdVm1atQqL26KMhcat0331hyTd1pYqIpI+SuAyJq9TI+PGhJfDssxt/rh/8IHQ1fvRR6CKeM6fx58wWK1bAtdeGEiwnnRR3NEGPHmFc3FNPhSLNlcrKwhquw4dn56LRhYVhAe7Jk/+7ragotC4OHBhbWCIiTY6SuAzp0SPMdMxkqZFNm0KB3+OPD9dPh+HD4eWXwz/poUPDYPWm4Fe/ColcpkuK1OUnPwmTTH760/B6QkjoEonsbdUaODAkbJVdqrNmwYwZ2RuviEiuUhKXQcOHhxmfK1Zk5nrFxWHlhXT/89xrL3jzzbAG67HHhkQxl334Ifztb3DeebDnnnFH812tWsFNN8HcuaE7G0Jy1L07HHdcvLHVprAQZs4Mt/HjoWVLOPPMuKMSEWlalMRl0LBhYeWEZ5/NzPWKisIMy2HD0n/uPn3CwPqDDgr1wX7969wtQXLJJaH0ym9+E3ck1TvhBDjiCPjlL0PC+cQTYSxcixZxR1azM88Midvtt8O994Yu6s6d445KRKRpURKXQfvtF/6RZWJcXEkJPPkknHNOGBMXhU6dQmvfOeeEBOPcc//b5ZcriovD7dprQ+tWNqosObJyJRx1VBgTN3Zs3FHVrnPnkLjdfnuY4KCuVBGR9Ivo37tUp1mz/5YaqaiIdu3Ie+8NBVej/mffsmUo7tq/fxhX9skncOCB0V4znR54ICwc/+Mfxx1J7XbfPUws+cc/woeB3XaLO6K6FRaGn2/v3qElUURE0ktJXIYNHw733x8Ge0e1pJN76Erdf3/YeedorlGVGVx3HfTtGwbgv/Za9NdMl/btw9JVLVvGHUndfv3rUET3pz+NO5LUHHFE+B08/fTwAUZERNJLSVyGHXNM+DplSnRJ3JtvhoHwda29mW5jx2Z/N18u69q1+sK/2apZs9xK6EVEco3GxGVY9+5hPdIoS40UFYWB+qedFt01REREJF5K4mIwbFhYRaFqRft0WbMmdNeeeipssUX6zy8iIiLZQUlcDCpLjTzzTPrP/fDDsHq1ZgOKiIg0dUriYrDPPtClSzSlRoqKYPvtQ/02ERERabqUxMWgWbMwwaGy1Ei6zJ8PL70UJhdk09JRIiIikn6RJXFm1tvMXjCzD8xsjpldlNx+avJxhZkNrrL/UWb2jpnNTn49vIbzXmdmX5nZrORteFTfQ5SGD4elS+Gdd9J3zgkTQu25c85J3zlFREQkO0VZYqQMuNTdZ5hZB+AdM3sWeB84Gbh9s/2/Br7n7ovMbADwNLB1Dee+xd3/FFXgmXDMMaG1rLg4dK82Vnl5SOKOOSasaSoiIiJNW2Qtce6ecPcZyfurgbnA1u4+193/p9qVu89090XJh3OANmbWKqr44ta1K+y7b/pKjTz7LHz1lSY0iIiI5IuMjIkzs37AIODNFA/5PjDD3TfU8PyPzOw9Mysysy1ruOb5ZjbdzKYvXbq0/kFnwLBhoTDv1183/lxFRWGyxPe+1/hziYiISPaLPIkzs/bAw8DF7r4qhf13A/4A/KCGXf4BbAcMBBLATdXt5O53uPtgdx/crVu3hoQeueHDwxJZjS01smwZPP44nH02tGqybZciIiJSVaRJnJm1ICRwk9z9kRT23wZ4FDjH3T+pbh93X+zu5e5eAdwJ7JvOmDNp772hW7fGlxqZPBk2btSSVyIiIvkkytmpBowD5rr7zSns3wl4CrjS3V+tZb+eVR6eRJgokZMKCuDYY+Hpp8PEhIYqKgoJ4Z57pi82ERERyW5RtsQNBUYBh1ctB2JmJ5nZQmB/4Ckzezq5/4+A7YFrq+zfHcDM7qpSjuSPyTIk7wGHAT+N8HuI3LBhYUzc9OkNO37mTJg1SxMaRERE8k1kJUbc/RWgppKzj1az/2+B39ZwrvOq3B+VlgCzxNFHhxa54mLYb7/6H19UFMbBnXFG+mMTERGR7KUVG2LWpUtI3hoyLm79epg0CU46Cbasdo6uiIiINFVK4rLAsGGhO7W+lVAefxxWrFBXqoiISD5SEpcFKkuNPP103ftWVVQEffrA4dUuUCYiIiJNmZK4LDBoEHTvXr8u1S++CKs0jBkDzZpFFpqIiIhkKSVxWaCgIHSp1qfUyMSJofVuzJhIQxMREZEspSQuSwwbBsuXw1tv1b1vRQWMHx+6Ufv3jz42ERERyT5K4rJE1VIjdXnpJfjsM01oEBERyWdK4rLEllvC/vunNi6uqAg6doSTT44+LhEREclOSuKyyLBh8M47sHhxzfusXAkPPRSK+7Zpk7nYREREJLsoicsiw4eHr//5T8373H9/KPKrrlQREZH8piQuiwwcCFttVfu4uKIiGDAABg+ueR8RERFp+pTEZRGz0KX6zDNQVva/z7//fpi9WlgY9hUREZH8pSQuywwbFpbSevPN/31u/Hho3hzOPjvzcYmIiEh2URKXZY46KqzAsHmX6qZNcM89cMIJ0K1bPLGJiIhI9lASl2U6dYIDDvjfUiNPPQVLl2pCg4iIiARK4rLQ8OEwcyYkEv/dVlQEPXvCMcfEF5eIiIhkDyVxWWjYsPC1stRIIhFa5s45J4yJExEREVESl4X22AN69frvuLh77oHychg7Nt64REREJHsoictCVUuNbNoUulKHDoWddoo7MhEREckWSuKy1LBhYYmtm2+Gjz7ShAYRERH5LnP3uGOI3ODBg3369Olxh1EvK1dC165QUAAtWoRxcR06xB2ViIiIRM3M3nH3OtdmUktclurYMXShbtwIp52mBE5yU3lFedwhiIg0WUristhxx4Wv6kqVXLSpfBO7/G0Xjp98PKs3rI47HBGRJkdJXBa78EJ48snQIieSa56a9xTzls/jqXlPcfCEg1m0elHcIYmINClK4rJY27ahNU6L3UsuKppZRM/2PXnyjCeZv3w++921H7MXz447LBGRJkNJnIikXWJ1ginzpnDOnudw3I7HMW3sNCq8ggPHH8hznz4Xd3giIk2CkjgRSbt73ruHci9n7MBQoXrgVgN549w36NOxD8MmDWPCrAnxBigi0gQoiRORtHJ3imYWMbT3UHbq+t8K1b079uaVsa9waL9DGfv4WK578TryocSRiEhUlMSJSFq9vvB1Plr2EYWD/ndadcfWHZly5hTGDBzDr176FWMfH8vG8o0xRCkikvu0nLqIpFXRzCLatWjHqbueWu3zLZq1oOiEIrbttC3XvngtX676kodPe5hOrTtlNlARkRynljgRSZvSjaU8MOcBTtvtNDq0qrlCtZnxi0N+wd0n3s20z6dxYNGBfLHyiwxGKiKS+5TEiUjaPPTBQ5RuLK22K7U6o/YcxX/O/g8LVy1kyF1DmJmYGXGEIiJNh5I4EUmboplF7NB5B4b2Tr1C9eH9D+fVwldp0awFB40/iCnzpkQYoYhI06EkTkTSYt6yeUz7YhqFgwqxelao3q37brxx7hvs1HUnTrjvBG6ffntEUYqINB1K4kQkLSbMmkCBFXDOnuc06PieHXry0piXOGb7Y/jhUz/kyueupMIr0hyliEjToSRORBqtvKKcCe9OYNj2w+jVoVeDz9O+ZXseH/k4P9z7h/zh1T9w1iNnsaFsQxojFRFpOlRiREQa7ZlPnmHR6kX8ZdhfGn2u5gXN+ftxf6f/lv254rkr+GrVVzw28jE6t+mchkhFRJoOtcSJSKMVzSqia9uuHL/j8Wk5n5lx+dDLuf/79/PmV29ywLgD+HTFp2k5t4hIU6EkTkQa5eu1X/P4h48zao9RtGzWMq3nPn3A6Uw9ZypL1y5lyF1DeOurt9J6fhGRXBZZEmdmvc3sBTP7wMzmmNlFye2nJh9XmNngzY65yszmm9lHZnZMDeftb2ZvJvd7wMzS+19DROpl0nuT2FSxKeXacPV1YJ8Dea3wNTq06sChEw7lsQ8fi+Q6IiK5JsqWuDLgUnffFRgCXGhmuwLvAycDL1fdOfncSGA34Fjg72bWrJrz/gG4xd23B1YA50b3LYhIbdydcTPHMbjXYAZ0HxDZdXbquhOvn/s6e/TYg5MfOJnb3rwtsmuJiOSKyCY2uHsCSCTvrzazucDW7v4sUF0dqRHA/e6+AfjMzOYD+wKvV+5g4aDDgTOTmyYC1wH/iOr7iNP0RdMZ9eioSBYI79S6E8VnFdO9Xfe0nztXLF+3nKPvOZoV61ek/dwtClrw84N+zqg9R6X93NlkRmIGs5fM5u/D/x75tbq3687zo5/n7EfO5qL/XMTNr99Ms4LqPueJiETntmNv47gdj4s7DCBDs1PNrB8wCHizlt22Bt6o8nhhcltVXYBv3L2sln0qr3k+cD5Anz596h90FnhpwUt8+PWHjBwwkuYF6XupyivKue/9+5gwawKXD708befNNU9+/CTvJN7h5F1Opm2Ltmk994dff8g5j53DZ998xi8O/kW9i9/miqKZRbRu3pozdj8jI9dr26ItD576IDe9fhOzl8zOyDVFRKrq2rZr3CF8K/IkzszaAw8DF7v7qqivV8nd7wDuABg8eLBn6rrpVFJaQqtmrZh88uS0JwFfrvqSoplFXHbAZU02wahL8fxitmq/FQ+e+iAFlt6RBRvLN/J///4/fvniL/nsm8+4/fjb0z7oP27rNq1j8vuTOXmXk+nUulPGrtusoFlef/gQEakU6exUM2tBSOAmufsjdez+FdC7yuNtktuqWgZ0MrPmtezTZCRKE/Ts0DOSJKtwYCEfLfuI1xe+XvfOTVBZRRlPz3+aY7c/Nu0JHEDLZi2ZMGICvzzkl0yYNYHjJh/HyvUr036dOD324WN8s/4bCgdGM6FBRERqF+XsVAPGAXPd/eYUDnkCGGlmrcysP7AD8J16Au7uwAvAKclNo4HH0xd1dikpLWGr9ltFcu5TdzuVdi3aUTSzKJLzZ7u3vnqLFetXMHz74ZFdw8y47tDrGD9iPC8ueJEDxx/Ilyu/jOx6mTZ+1nj6duzLYf0PizsUEZG8FGVL3FBgFHC4mc1K3oab2UlmthDYH3jKzJ4GcPc5wL+AD4D/ABe6ezmAmU0xs8q1fK4ALklOfOhCSBSbpJLSEnq27xnJudu3bM/pu53OA3MeoHRjaSTXyGZT5k2hmTXjqO2OivxaYwaOofisYr5Y+QVDxg1hVsmsyK8Ztc+/+ZznPn2OsQPHRtKSKSIidavzr6+Z7WhmU83s/eTjPczsmrqOc/dX3N3cfQ93H5i8TXH3R919G3dv5e493P2YKsdc7+7buftO7l5cZftwd1+UvP+pu+/r7tu7+6nJ2axNUqI0EVlLHEDhoEJKN5by0AcPRXaNbFU8v5j9e++fsbFcR257JK+MfYUCK+Cg8Qfxn/n/ych1ozLx3YlASFBFRCQeqXyEvhO4CtgE4O7vEeq5SYQ2lG1g+brlkbXEARzQ+wB27LJj3nWplpSWMCMxI9Ku1Ors3mN33jzvTbbvvD3HTz6eO9+5M6PXT5cKr2D8rPEcse0R9O3UN+5wRETyVipJXFt333ytm7Jq95S0WbxmMUCkLXFmxtiBY5n2xTTmLZsX2XWyTWUr2LAdhmX82r069OLlMS9z1HZHcf6T53P11Kup8IqMx9EYLy54kQXfLNCEBhGRmKWSxH1tZtsBDmBmp5As4ivRKSktAaBnh+ha4gDO2fMcCqyACbMmRHqdbDJl3hR6tu/Jnj32jOX6HVp14N9n/Jvz9zqf37/ye85+5Gw2lOXOqICimUV0at2JE3c+Me5QRETyWipJ3IXA7cDOZvYVcDHw/6IMSiCxOuTJUbbEQWgZGrb9MCa8O4HyivJIr5UNyirKeOaTZxi2/bBY6+M1L2jOP4//J78/4vfc9/59HH3v0Sxftzy2eFL1zfpveHjuw5w54EzatGgTdzgiInmtziQuOZHgSKAbsLO7H+juCyKPLM992xIX4Zi4SoWDClm0ehHPfPJM5NeK2+tfvs7KDSsZvkNmx8NVx8y48sArmXzyZN5Y+AZDi4by2YrP4g6rVve/fz/ry9ZHtti9iIikLpXZqReZ2RbAWuAWM5thZkdHH1p+S5QmMCwja5sev+PxdG3blaJZTX+CQ/H8YpoXNOfIbY+MO5RvnbH7GTw76lkWly5myLghvP3V23GHVKOimUXs0WMP9uq5V9yhiIjkvVS6UwuTy2UdTajLNgq4IdKohJLSErq27UqLZi0iv1bLZi0ZtccoHv/wcb5e+3Xk14tT8fxihvYeSsfWHeMO5TsO7nswr537Gm1btOWQCYfw+IfZV8N69uLZvL3obcYOHJu3S7WJiGSTVJK4yr/Ww4G7k0V59Rc8YlHXiNtc4aBCNlVsYtJ7kzJ2zUxbtHoRs0pmMWz7zM9KTcXOXXfmjXPfYED3AZz0wEn85c2/xB3Sd4yfNZ4WBS04a/ez4g5FRERILYl7x8yeISRxT5tZByC3aiLkoJLSkshnplY1oPsA9um1D+NmjiOsbtb0xFlaJFU92vfghdEvcMJOJ/CT//yES56+JCtKkGws38g9793DCTudQLd23eIOR0RESC2JOxe4EtjH3dcCLYGxkUYlJFZntiUOQmvc7CWzmZGYkdHrZsqUeVPYusPW7N5997hDqVW7lu14+LSH+cm+P+GWN27h1AdPZd2mdbHG9NTHT/H12q81oUFEJIvUmMSZ2V5mthcwMLlp2+TjvkDzDMSWt9w90nVTazJywEhaN2/dJFdw2FS+iWc/fTb20iKpalbQjD8P+zO3HHMLj859lCPuPoKla5bGFk/RrCJ6dejF0dtpTpOISLaoLRm7qZbnHDg8zbFI0vJ1y9lUsSnjLXGdWnfi+7t8n8nvT+ZPR/+pSdUBe+3L11i1YVVWlBapj4uHXEyfjn0465Gz2H/c/hSfVcwOXXbIaAyLVi9iyrwpXDH0CpoX6PObiEi2qLElzt0Pq+WmBC5CmawRt7nCQYV8s/4bHvvwsYxfO0qVpUWO2PaIuEOpt5N3OZnnz3melRtWsv+4/Xn1i1czev173r2HCq9g7ECNohARySapjInDzAaY2Wlmdk7lLerA8lmiNDOrNVTn0H6H0q9TvyZXM27KvCkc1Ocgtmi1RdyhNMj+vffnjXPfoHObzhxx9xE8OOfBjFzX3SmaVcRBfQ7KeAugiIjULpViv78E/pK8HQb8ETgh4rjyWqbWTa1OgRUwZs8xTP10Kp9/83nGrx+FhasWMnvJ7KwtLZKq7Tpvx2vnvsbgXoM57aHTuPHVGyOfSfzal6/x8bKPNaFBRCQLpdISdwpwBFDi7mOBPYHsqpTaxGRq3dSajBk4BoCJ706M5frpVllaJNfGw1Wna9uuPHfOc5y666lc/tzl/GjKjyirKIvsekUzi2jfsj2n7HpKZNcQEZGGSSWJW+fuFUBZcvmtJUDvaMPKbyWlJbRt0ZYOLTvEcv2+nfpyxLZHMH7W+KyoUdZYU+ZNofcWvdm1265xh5IWrZu35v5T7ueyAy7j79P/zkkPnMSajWvSfp3SjaU8MOcBTt/tdNq3bJ/284uISOOkksRNN7NOwJ3AO8AM4PUog8p3las1xFkKo3BgIQu+WcCLC16MLYZ02Fi+kec+fS5nSoukqsAK+ONRf+Rvw//GlHlTOGTCId92w6fLg3MeZM2mNepKFRHJUrXViTsZwN0vAMzd/wkcBYxOdqtKROKoEbe5E3c+kU6tO+V8zbhXv3iV1RtXN4mu1OpcsM8FPD7yceZ+PZchdw3hg6UfpO3cRbOK2KnLTuy/zf5pO6eIiKRPbS1x11S5PxXA3Re4+3vRhiSZXje1Om1atOHMAWfy8NyH+Wb9N7HG0hjF84tpUdCCw/s33ao4x+94PC+PeZkN5Rs4YNwBvPDZC40+58fLPuaVL16hcFBhk2rBFBFpSmpL4qyG+xKxktKS2JM4CDXj1pet5/737487lAabMm8KB/c9mA6t4hlfmCl799qbN859g14denHMvcdw73v3Nup8E2ZNoJk1Y9Qeo9IUoYiIpFttSVwbMxtkZnsDrZP396qyHJdEYN2mdXyz/pvYu1MB9uq5F3v02CNnu1S/WPkFc5bOyfnSIqnq26kvrxa+ytA+Qxn16Ch++/JvG1SCpKyijInvTmTYDsNiKXMjIiKpqS2JSwA3A38CSpL3b0re/hR9aPlp8ZrFQHzlRaoyMwoHFvL2oreZvXh23OHUW/G8YqBplBZJ1ZZttuQ/Z/2Hs/c4m1+88Av+79//x6byTfU6xzOfPMOi1YsoHKgJDSIi2UzLbmWZyhpx2dICctYeZ9GioAXjZ42PO5R6K55fTN+Ofdm5685xh5JRrZq34u4T7+aag65h3MxxHH/f8azasCrl44tmFtGtbTeO2/G4CKMUEZHGSmnZLcmcyjIR2dASB6G47IidR3DPe/ewsXxj3OGkbEPZBp779DmG7zA8Lwfmmxm/Ofw3jDthHM9/9jwHjT+IhasW1nnc0jVLeeKjJxi1xyhaNmuZgUhFRKShlMRlmW+X3MqCMXGVxg4cy9drv+bJj5+MO5SUvfLFK6zZtCZvxsPVpHBQIU+d+RSfrfiMIXcN4d2Sd2vdf9LsSWyq2MTYQaoiJCKS7ZTEZZlEaQLD6NauW9yhfOvo7Y6mV4deOTXBoXh+MS2btWzSpUVSdfR2RzNt7DQADhp/EM988ky1+7k7RTOL2HfrfRnQfUAmQxQRkQaordjvXrXdMhlkPikpLaF7u+40L2gedyjfal7QnNF7jqZ4fjGLVi+KO5yUTJk3hUP6HkK7lu3iDiUr7LnVnrxx3hv037I/wycNrzYhfyfxDrOXzNaEBhGRHFFbS1zlTNS/AW8CdxCW3nozuU0ikA2FfqszduBYKryCu9+9O+5Q6rTgmwXM/Xpu3nelbm6bLbZh2thpHLHtEZz7xLlc8/w13ylBUjSziNbNWzNywMgYoxQRkVTVOTuVUGpkL3cf7O57A4OArzIVYL4pKS3JmpmpVe3QZQcO6nMQRTOLGlR7LJPysbRIqrZotQVPnvEk5w46l+unXc+oR0exoWwD6zatY/LsyZyy6yl0bN0x7jBFRCQFqYyJ28ndvy0S5u7vA7tEF1J+S6zOzpY4CIPk5y2fx2tfvhZ3KLUqnl9M/0792bHLjnGHkpVaNGvBnd+7k98c9hsmzZ7EsZOOpWhmESs3rFRXqohIDkkliXvPzO4ys0OTtzsBrZ8agQqvYPGaxVk1M7WqU3Y9hfYt22f1BIf1ZeuZ+tnUvC0tkioz45qDr+Gek+7h1S9e5UfFP6J/p/4c0u+QuEMTEZEUpZLEjQXmABclbx8kt0maLVu7jLKKsqxtiWvfsj2n73Y6D8x5gNKNpXGHU61pn09j7aa1Gg+XorP3OJtnRj1Dt7bduHjIxRSYJqyLiOSKOv9iu/t6d7/F3U9K3m5x9/WZCC7fZGONuM0VDipkzaY1PDjnwbhDqdaUeVNo1awVh/U/LO5Qcsah/Q6l5Gcl/GS/n8QdioiI1EOdSZyZDTWzZ83sYzP7tPKWieDyTaI0LLmVrS1xAPtvsz87ddmJolnZ2aVaPL+YQ/sdStsWbeMOJaeoBU5EJPek8pd7HHAzcCCwT5WbpNm3LXFZODu1kplROKiQV754hY+XfRx3ON/x6YpP+WjZR+pKFRGRvJBKErfS3YvdfYm7L6u8RR5ZHkqszv6WOIBRe4yimTVj/MzxcYfyHSotIiIi+SSVJO4FM7vRzPbXig3RKiktoX3L9rRv2T7uUGrVs0NPhu0wjInvTqSsoizucL5VPL+Y7bbcjh267BB3KCIiIpFLZW2n/ZJfB1fZ5oAWpUyzbF2toTqFAwt58uMneXr+0xy343Fxh8P6svU8/9nznLfXeXGHIiIikhF1JnHJVRvqzcx6A3cDPQhJ3x3u/mcz6ww8APQDFgCnufsKM7sMOKtKXLsA3dx9+WbnnQAcAqxMbhrj7rMaEmO2KSktyeqZqVUdt+NxdGvbjaJZRVmRxL204CXWla3TeDgREckbKa2ybmbHAbsBrSu3ufuv6zisDLjU3WeYWQfgHTN7FhgDTHX3G8zsSuBK4Ap3vxG4MXm97wE/3TyBq+Iyd38oldhzSaI0wZ499ow7jJS0bNaSUXuM4ra3bmPpmqV0a9ct1nimzJtC6+atObTfobHGISIikimplBj5J3A68GPAgFOBvnUd5+4Jd5+RvL8amAtsDYwAJiZ3mwicWM3hZwD31R1+05JLLXEAYweNpayijHvfuzfuUCieX8xh/Q6jTYs2cYciIiKSEalMbDjA3c8BVrj7r4D9gXotSmlm/YBBwJtAD3dPJJ8qIXS3Vt23LXAs8HAtp7zezN4zs1vMrFUN1zzfzKab2fSlS5fWJ9xYrN20llUbVuXMmDiAAd0HsO/W+zJu5jjcPbY45i+fz7zl8zQrVURE8koqSdy65Ne1ZtYL2ASk3FxkZu0JCdnF7r6q6nMe/vNv/t//e8CrtXSlXgXsTKhV1xm4orqd3P0Odx/s7oO7dYu3qy8VuVAjrjqFAwuZs3QO0xdNjy2GytIiGg8nIiL5JJUk7kkz60QYrzaDMBlhcionN7MWhARukrs/kty82Mx6Jp/vCSzZ7LCR1NKVmuymdXffAIwH9k0llmyXKzXiNjdywEhaN2/N+Fnx1YybMn8KO3Tege06bxdbDCIiIpmWytqpv3H3b9z9YcJYuJ3d/dq6jjMzI6z2MNfdb67y1BPA6OT90cDjVY7pSJh5+jg1qJIAGmE83ft1xZILcmHd1Op0bN2RU3Y9hcmzJ7Nu07q6D0izdZvW8eKCF9WVKiIieadeCya6+wZ3X1n3ngAMBUYBh5vZrORtOHADcJSZzQOOTD6udBLwjLuvqXoiM5uS7MoFmGRms4HZQFfgt/X5HrJVLqybWpPCgYWs3LCSRz98NOPXfnHBi6wvW6+uVBERyTsplRhpCHd/hTCbtTpH1HDMBGBCNduHV7nfJIsMl5SW0Mya0bVt17hDqbdD+h1C/079KZpZxJm7n5nRa0+ZN4U2zdtwSL9DMnpdERGRuNWrJU6ik1idoHu77jQraBZ3KPVWYAWMGTiGqZ9NZcE3CzJ2XXdnyvwpHN7/cFo3b133ASIiIk1IKnXizMzONrNrk4/7mFmTmEyQTUrWlOTczNSqRu85GsOYMGtCxq45b/k8Pl3xqcbDiYhIXkqlJe7vhNpwZyQfrwb+FllEeaqktCQnx8NV6tupL0dueyTjZ42nwisyck2VFhERkXyWShK3n7tfCKwHcPcVQMtIo8pDidUJtmqXu0kcwLmDzuWLlV9wUfFFlFeUR369KfOnsHPXnem/Zf/IryUiIpJtUkniNplZM5JFec2sG5CZppY8UV5RzpI1S3K6OxXg1N1O5ZIhl/DXt//K9//1fdZuWhvZtdZsXMNLC15SK5yIiOStVJK424BHge5mdj3wCvC7SKPKM1+v/ZpyL8/p7lQIExxuOuYmbjv2Np746AkOm3gYS9ZsXss5PV5Y8AIbyjcoiRMRkbyVSrHfScDlwO+BBHCiuz8YdWD5JFcL/dbkx/v9mEdPf5TZi2cz5K4hfPT1R2m/RvG8Ytq2aMvBfQ9O+7lFRERyQSqzU/sAa4F/E1ZbWJPcJmmSy4V+azJi5xG8OOZF1mxaw/7j9mfa59PSdu7K0iJH9D+CVs1bpe28IiIiuSSV7tSngCeTX6cCnwLFUQaVb75ticvxMXGb23frfXnj3Dfo3q47R95zJPe/f39azvvRso9Y8M0ClRYREZG8lkp36u7uvkfy6w6EBedfjz60/JFY3fRa4ir137I/r537GkO2GcIZD5/BDa/cgLs36pwqLSIiItKAFRvcfQawXwSx5K2S0hK2aLUFbVu0jTuUSHRu05lnzn6GMwacwVVTr+KHT/6QsoqyBp9vyvwp7NptV/p26pvGKEVERHJLnWunmtklVR4WAHsBiyKLKA8lShNNshWuqlbNW3HvyffSr1M/fv/K7/ly1Zc8cMoDdGjVoV7nKd1Yysufv8yP9/1xRJGKiIjkhlRa4jpUubUijI0bEWVQ+aaktKTJzEytTYEV8Lsjfsftx9/OM588wyETDmHR6vp9Hnj+s+fZWL5R4+FERCTv1doSlyzy28Hdf5ahePJSojTB3j33jjuMjDl/7/PpvUVvTn3wVIbcNYQpZ01hQPcBKR1bPK+Y9i3bc2CfAyOOUkREJLvV2hLn7uXA0AzFkrfypSWuqmE7DGPa2GmUVZQxtGgoUz+dWucxlaVFjtz2SFo208pvIiKS32pM4sysspVulpk9YWajzOzkyluG4mvySjeWUrqxtMmPiavOoJ6DeOO8N+jTsQ/HTjqWibMm1rr/3K/n8sXKLzQrVUREhNpb4t5Kfm0NLAMOB76XvB0fcVx5o6nWiEtVn459eGXsKxza71DGPD6GX734qxpLkEyZNwVQaRERERGofUycAbj72AzFkpeaco24VHVs3ZGnznyK8/99Pte9dB0LVi7g9uNv/58u0+L5xQzoPoDeHXvHFKmIiEj2qC2J67ZZeZHvcPebI4gn7zS1dVMbqmWzlowfMZ7+nfpz3UvX8eXKL3n4tIfp2LojAKs3rGba59O4eMjF8QYqIiKSJWrrTm0GtOe7JUaq3iQNmuK6qQ1lZvzy0F8yYcQEXvr8JYYWDeWLlV8AMPWzqWyq2KTSIiIiIkm1tcQl3P3XGYskT5WUltC8oDld2naJO5SsMXrgaLbZYhtO/tfJDLlrCE+d+RTF84rp0LIDQ3trsrSIiAjU3hJnGYsijyVKE/Ro14MCq/cKaE3aEdsewauFr9K8oDkHjT+Ih+Y+xFHbHUWLZi3iDk1ERCQr1JY5HJGxKPJYSWlJ3s5MrcuA7gN447w32KHLDixft1yzUkVERKqosTvV3ZdnMpB8lVid0GzLWvTq0IuXx7zMv+b8i7P3ODvucERERLKG+vBilo+rNdRXh1YdOHevc2nVvFXcoYiIiGSNOpM4M/tDKtuk/soqyliyZolmpoqIiEi9pdISd1Q12zQ4KQ2WrlmK42qJExERkXqrcUycmf0/4AJgOzN7r8pTHYDXog4sH1QW+lVLnIiIiNRXbXXiJgPFwO+BK6tsX61JD+lRWehXs1NFRESkvmrsTnX3le6+APgzsNzdP3f3z4EyM9svUwE2ZWqJExERkYZKZUzcP4DSKo9Lk9ukkRKrteSWiIiINEwqSZy5u1c+cPcKau+GlRSVlJbQqXUnWjdvHXcoIiIikmNSSeI+NbOfmFmL5O0i4NOoA8sHidKEZqaKiIhIg6SSxP0QOAD4ClgI7AecH2VQ+aKktERdqSIiItIgdXaLuvsSYGQGYsk7idIE+22tOSIiIiJSf6ms2LCjmU01s/eTj/cws2uiD61pc3ctuSUiIiINlkp36p3AVcAmAHd/D7XMNdrqjatZu2mtulNFRESkQVJJ4tq6+1ubbSuLIph8UlkjToV+RUREpCFSSeK+NrPtAAcws1OARKRR5QHViBMREZHGSCWJuxC4HdjZzL4CLibMWK2VmfU2sxfM7AMzm5MsTYKZdTazZ81sXvLrlsnth5rZSjOblbxdW8N5+5vZm2Y238weMLOWqX6z2eTbljiNiRMREZEGqDWJM7NmwAXufiTQDdjZ3Q9MLr9VlzLgUnffFRgCXGhmuxLWYZ3q7jsAU/nuuqzT3H1g8vbrGs77B+AWd98eWAGcm0IsWady3VS1xImIiEhD1JrEuXs5cGDy/hp3X53qid094e4zkvdXA3OBrYERwMTkbhOBE1M9p5kZcDjwUEOOzyYlpSW0KGhB5zad4w5FREREclAqy2fNNLMngAeBNZUb3f2RVC9iZv2AQcCbQA93rxxTVwL0qLLr/mb2LrAI+Jm7z9nsVF2Ab9y9cmLFQkJiWN01zydZlLhPnz6phpoxidIEW7XfipCXioiIiNRPKklca2AZoQWskgMpJXFm1h54GLjY3VdVTVrc3c2scl3WGUBfdy81s+HAY8AOqVyjOu5+B3AHwODBg72O3TOupLREM1NFRESkwWpN4pJj4pa5+88acnIza0FI4CZVablbbGY93T1hZj2BJQDuvqryOHefYmZ/N7Ou7v51lVMuAzqZWfNka9w2hOXAck5idYL+W/aPOwwRERHJUamMiRvakBMnx6+NA+a6+81VnnoCGJ28Pxp4PLn/VsljMLN9k7Et2yweB14ATtn8+Fyj1RpERESkMVLpTp3VwDFxQ4FRwGwzm5XcdjVwA/AvMzsX+Bw4LfncKcD/M7MyYB0wMpm0YWZTgPPcfRFwBXC/mf0WmElIFHPKpvJNLF27VDNTRUREpMEiGxPn7q8ANY3aP6Ka/f8K/LWGcw2vcv9TYN/aQ85uS9YsAVQjTkRERBquziTO3cdmIpB8ohpxIiIi0lh1rthgZtuY2aNmtiR5e9jMtslEcE2V1k0VERGRxkpl2a3xhMkIvZK3fye3SQNp3VQRERFprFSSuG7uPt7dy5K3CYQluKSBKlvierTrUceeIiIiItVLJYlbZmZnm1mz5O1sNiv9IfWTKE3QuU1nWjVvFXcoIiIikqNSSeIKCWVASoAEoRSIJjs0gmrEiYiISGOlMjv1c+CEDMSSN0pKSzQeTkRERBolldmpE82sU5XHW5pZUaRRNXGJ0oRmpoqIiEijpNKduoe7f1P5wN1XAIMii6iJc/fQEtdOLXEiIiLScKkkcQVmtmXlAzPrTGorPUg1Vm5Yyfqy9WqJExERkUZJJRm7CXjdzB5MPj4VuD66kJq2yvIiGhMnIiIijZHKxIa7zWw6/1079WR3/yDasJquykK/mp0qIiIijZFSt2gyaVPilgZqiRMREZF0SGVMnKRRojTZEqcxcSIiItIISuIyrKS0hFbNWtGxVce4QxEREZEcpiQuwyprxJlZ3KGIiIhIDlMSl2FarUFERETSQUlchiVWJ5TEiYiISKMpicuwktISlRcRERGRRlMSl0EbyzeybN0ytcSJiIhIoymJy6DFpYsBFfoVERGRxlMSl0GVNeLUEiciIiKNpSQugypXa1ChXxEREWksJXEZVLluqlriREREpLGUxGVQZUtcj3Y9Yo5EREREcp2SuAxKlCbo2rYrLZq1iDsUERERyXFK4jJINeJEREQkXZTEZVCiVKs1iIiISHooicugktISzUwVERGRtFASlyHuTklpCVu1U0uciIiINJ6SuAxZsX4FG8s3qiVORERE0kJJXIZUlhfRmDgRERFJByVxGVJZ6FezU0VERCQdlMRliFriREREJJ2UxGVIojTZEqcxcSIiIpIGSuIypKS0hDbN29ChZYe4QxEREZEmQElchiRKE/Ts0BMzizsUERERaQKUxGVISWmJxsOJiIhI2iiJy5DE6oRmpoqIiEjaRJbEmVlvM3vBzD4wszlmdlFye2cze9bM5iW/bpncfpaZvWdms83sNTPbs4bzTjCzz8xsVvI2MKrvIZ3UEiciIiLpFGVLXBlwqbvvCgwBLjSzXYErganuvgMwNfkY4DPgEHffHfgNcEct577M3Qcmb7Mi+w7SZH3ZelasX6GWOBEREUmbyJI4d0+4+4zk/dXAXGBrYAQwMbnbRODE5D6vufuK5PY3gG2iii3TFpcuBlQjTkRERNInI2PizKwfMAh4E+jh7onkUyVAj2oOORcoruWU1ye7Xm8xs1Y1XPN8M5tuZtOXLl3aiOgbTzXiREREJN0iT+LMrD3wMHCxu6+q+py7O+Cb7X8YIYm7ooZTXgXsDOwDdK5pP3e/w90Hu/vgbt26Ne6baCSt1iAiIiLpFmkSZ2YtCAncJHd/JLl5sZn1TD7fE1hSZf89gLuAEe6+rLpzJrtp3d03AOOBfaP8HtJB66aKiIhIukU5O9WAccBcd7+5ylNPAKOT90cDjyf37wM8Aoxy949rOW9lAmiE8XTvpz34NCspLcEwurWLt0VQREREmo7mEZ57KDAKmG1ms5LbrgZuAP5lZucCnwOnJZ+7FugC/D25qkGZuw8GMLMpwHnuvgiYZGbdAANmAT+M8HtIi0Rpgu7tutO8IMoft4iIiOSTyLIKd3+FkGhV54hq9j8POK+Gcw2vcv/wtASYQaoRJyIiIummFRsyIFGaUBInIiIiaaUkLgNKSktUXkRERETSSklcxCq8InSntlNLnIiIiKSPkriILV+3nLKKMrXEiYiISFopiYtYZY04jYkTERGRdFISF7HK1RpU6FdERETSSUlcxCrXTVVLnIiIiKSTkriIfdsSpzFxIiIikkZK4iKWWJ2gXYt2tG/ZPu5QREREpAlREhexkjWqESciIiLppyQuYlpyS0RERKKgJC5iidUJzUwVERGRtFMSFzG1xImIiEgUlMRFaN2mdazcsFItcSIiIpJ2SuIiVFleRC1xIiIikm5K4iJUWehXs1NFREQk3ZTERUgtcSIiIhIVJXERSqxOtsRpTJyIiIikmZK4CJWUllBgBXRt2zXuUERERKSJURIXoURpgh7tetCsoFncoYiIiEgToyQuQqoRJyIiIlFREhehRGlCM1NFREQkEkriIlRSWsJW7dQSJyIiIumnJC4i5RXlLC5drJY4ERERiYSSuIgsW7eMci/XmDgRERGJhJK4iKhGnIiIiERJSVxEtFqDiIiIRElJXES0bqqIiIhESUlcRNQSJyIiIlFSEheRxOoEW7TagrYt2sYdioiIiDRBSuIiUrJGqzWIiIhIdJTERSSxOqEkTkRERCKjJC4iJaUlKi8iIiIikVESF5FEqVriREREJDpK4iJQurGU0o2laokTERGRyCiJi8Di0sWAyouIiIhIdJTERUCFfkVERCRqSuIioEK/IiIiEjUlcRFIrE62xGlMnIiIiEQksiTOzHqb2Qtm9oGZzTGzi5LbO5vZs2Y2L/l1y+R2M7PbzGy+mb1nZnvVcN69zWx2cr/bzMyi+h4aqqS0hOYFzenStkvcoYiIiEgTFWVLXBlwqbvvCgwBLjSzXYErganuvgMwNfkYYBiwQ/J2PvCPGs77D+D/qux7bGTfQQMlShP0aNeDAlNDp4iIiEQjsizD3RPuPiN5fzUwF9gaGAFMTO42ETgxeX8EcLcHbwCdzOw7/ZHJx1u4+xvu7sDdVY7PGiWlWnJLREREopWRpiIz6wcMAt4Eerh7IvlUCdAjeX9r4Msqhy1Mbqtq6+T22vapvOb5ZjbdzKYvXbq0cd9APSVKE5qZKiIiIpGKPIkzs/bAw8DF7r6q6nPJ1jSP4rrufoe7D3b3wd26dYviEjUqKS1hq3ZqiRMREZHoRJrEmVkLQgI3yd0fSW5eXNlNmvy6JLn9K6B3lcO3SW6r6qvk9tr2iVV5RTlL1ixRS5yIiIhEKsrZqQaMA+a6+81VnnoCGJ28Pxp4vMr2c5KzVIcAK6t0uwJhnB2wysyGJM9/TpXjs8LStUup8AqNiRMREZFINY/w3EOBUcBsM5uV3HY1cAPwLzM7F/gcOC353BRgODAfWAuMrTyRmc1y94HJhxcAE4A2QHHyljVUI05EREQyIbIkzt1fAWqq4XZENfs7cGEN5xpY5f50YEAaQoyEVmsQERGRTFAhszTTuqkiIiKSCUri0kwtcSIiIpIJSuLSLLE6QafWnWjdvHXcoYiIiEgTpiQuzUrWaLUGERERiZ6SuDRLrE5oZqqIiIhETklcmmndVBEREckEJXFp5O5h3VS1xImIiEjElMSlUenGUtZuWquWOBEREYmckrg0Uo04ERERyRQlcWmkGnEiIiKSKUri0kjrpoqIiEimKIlLI7XEiYiISKYoiUujktISWhS0oHObznGHIiIiIk2ckrg0SpQm2Kr9VphZ3KGIiIhIE6ckLo1U6FdEREQyRUlcGiVKEyovIiIiIhmhJC6NSkpL2KqdWuJEREQkekri0qSsooyla5aqJU5EREQyQklcmixZswTHNSZOREREMkJJXJqo0K+IiIhkkpK4NFGhXxEREckkJXFpkihNtsRpTJyIiIhkgJK4NKlsievRrkfMkYiIiEg+UBKXJonVCTq36Uyr5q3iDkVERETygJK4NClZo9UaREREJHOUxKVJYnVCM1NFREQkY5TEpYnWTRUREZFMUhKXBu4e1k1VS5yIiIhkiJK4NFi1YRXry9arJU5EREQyRklcGqhGnIiIiGSakrg00GoNIiIikmlK4tJA66aKiIhIpimJSwO1xImIiEimNY87gKbg9AGns1v33ejUulPcoYiIiEieUBKXBr069KJXh15xhyEiIiJ5RN2pIiIiIjlISZyIiIhIDlISJyIiIpKDlMSJiIiI5KDIkjgzKzKzJWb2fpVte5rZ62Y228z+bWZbJLefZWazqtwqzGxgNee8zsy+qrLf8KjiFxEREclmUbbETQCO3WzbXcCV7r478ChwGYC7T3L3ge4+EBgFfObus2o47y2V+7r7lEgiFxEREclykSVx7v4ysHyzzTsCLyfvPwt8v5pDzwDujyouERERkaYg02Pi5gAjkvdPBXpXs8/pwH21nONHZvZesrt2y5p2MrPzzWy6mU1funRpwyMWERERyUKZTuIKgQvM7B2gA7Cx6pNmth+w1t3fr+5g4B/AdsBAIAHcVNOF3P0Odx/s7oO7deuWjthFREREskZGV2xw9w+BowHMbEfguM12GUktrXDuvrjyvpndCTwZQZgiIiIiWS+jLXFm1j35tQC4BvhnlecKgNOoZTycmfWs8vAkoKYWOxEREZEmLcoSI/cBrwM7mdlCMzsXOMPMPgY+BBYB46sccjDwpbt/utl57jKzwcmHf0yWJ3kPOAz4aVTxi4iIiGQzc/e4Y4jc4MGDffr06XGHISIiIlInM3vH3QfXtZ9WbBARERHJQUriRERERHJQXnSnmtlS4PPkw67A1zGGIw2n1y436XXLTXrdcpNet9xV9bXr6+511kfLiySuKjObnko/s2QfvXa5Sa9bbtLrlpv0uuWuhrx26k4VERERyUFK4kRERERyUD4mcXfEHYA0mF673KTXLTfpdctNet1yV71fu7wbEyciIiLSFORjS5yIiIhIzlMSJyIiIpKD8iqJM7NjzewjM5tvZlfGHY+kxswWJNfMnWVmWj8ti5lZkZktMbP3q2zrbGbPmtm85Nct44xR/lcNr9t1ZvZV8n03y8yGxxmj/C8z621mL5jZB2Y2x8wuSm7Xey6L1fK61fs9lzdj4sysGfAxcBSwEHgbOMPdP4g1MKmTmS0ABru7ClhmOTM7GCgF7nb3AcltfwSWu/sNyQ9PW7r7FXHGKd9Vw+t2HVDq7n+KMzapmZn1BHq6+wwz6wC8A5wIjEHvuaxVy+t2GvV8z+VTS9y+wHx3/9TdNwL3AyNijkmkSXH3l4Hlm20eAUxM3p9I+GMlWaSG102ynLsn3H1G8v5qYC6wNXrPZbVaXrd6y6ckbmvgyyqPF9LAH5pknAPPmNk7ZnZ+3MFIvfVw90TyfgnQI85gpF5+ZGbvJbtb1SWXxcysHzAIeBO953LGZq8b1PM9l09JnOSuA919L2AYcGGy60dykIfxG/kxhiP3/QPYDhgIJICbYo1GamRm7YGHgYvdfVXV5/Sey17VvG71fs/lUxL3FdC7yuNtktsky7n7V8mvS4BHCV3jkjsWJ8eAVI4FWRJzPJICd1/s7uXuXgHcid53WcnMWhASgUnu/khys95zWa66160h77l8SuLeBnYws/5m1hIYCTwRc0xSBzNrlxz4iZm1A44G3q/9KMkyTwCjk/dHA4/HGIukqDIJSDoJve+yjpkZMA6Y6+43V3lK77ksVtPr1pD3XN7MTgVITte9FWgGFLn79fFGJHUxs20JrW8AzYHJet2yl5ndBxwKdAUWA78EHgP+BfQBPgdOc3cNos8iNbxuhxK6dRxYAPygyjgryQJmdiAwDZgNVCQ3X00YX6X3XJaq5XU7g3q+5/IqiRMRERFpKvKpO1VERESkyVASJyIiIpKDlMSJiIiI5CAlcSIiIiI5SEmciIiISA5SEicishkzK434/N3M7E0zm2lmB2323F1mtmuU1xeRpkElRkRENmNmpe7ePsLzjwSOdPfzorqGiDR9aokTkSbNzG4wswurPL7OzH5mZu3NbKqZzTCz2WY2oppjDzWzJ6s8/quZjUne39vMXjKzd8zs6c2qrVfu38/Mnk8uaD3VzPqY2UDgj8AIM5tlZm02O+ZFMxucvp+AiDRVSuJEpKl7ADityuPTktvWAye5+17AYcBNyeVw6pRc9/AvwCnuvjdQBFS3kshfgInuvgcwCbjN3WcB1wIPuPtAd1/XsG9LRPJd87gDEBGJkrvPNLPuZtYL6AascPcvk4nY78zsYMLSN1sDPYCSFE67EzAAeDaZ9zUDqlseZ3/g5OT9ewgtcCIiaaEkTkTywYPAKcBWhFY4gLMISd3e7r7JzBYArTc7rozv9lhUPm/AHHffP7KIRUTqoO5UEckHDwAjCYncg8ltHYElyQTuMKBvNcd9DuxqZq3MrBNwRHL7R0A3M9sfQveqme1WzfGvJa8LIWmclo5vRkQE1BInInnA3eeYWQfgK3ev7PacBPzbzGYD04EPqznuSzP7F/A+8BkwM7l9o5mdAtxmZh0Jf0tvBeZsdoofA+PN7DJgKTA27d+ciOQtlRgRERERyUHqThURERHJQUriRERERHKQkjgRERGRHKQkTkRERCQHKYkTERERyUFK4kRERERykJI4ERERkRz0/wFlRag9uAUI/wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "list_true = []\n",
    "list_flase = []\n",
    "list_i = []\n",
    "for i in range(1, 25):\n",
    "    from sklearn.ensemble import RandomForestClassifier\n",
    "    \n",
    "    rf_classifier = RandomForestClassifier(n_estimators=i, criterion='entropy')\n",
    "    rf_classifier.fit(X_train, y_train)\n",
    "\n",
    "    # Predicting the Test set results\n",
    "    y_pred = rf_classifier.predict(X_test)\n",
    "\n",
    "    # Making the Confusion Matrix\n",
    "    from sklearn.metrics import confusion_matrix\n",
    "\n",
    "    cm5 = confusion_matrix(y_test, y_pred)\n",
    "    from sklearn.metrics import accuracy_score\n",
    "\n",
    "    ac5 = accuracy_score(y_test, y_pred)\n",
    "    \n",
    "    list_i.append(i)\n",
    "    list_true.append(cm5[0][0])\n",
    "    list_flase.append(cm5[1][1]+200)\n",
    "plt.figure(figsize=(10,7))\n",
    "plt.plot(list_i, list_true, \"g\")\n",
    "plt.plot(list_i, list_flase, \"b\")\n",
    "plt.xlabel(\"value of i\")\n",
    "plt.ylabel(\"correct True and False\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e799cdb8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:01.275918Z",
     "start_time": "2021-05-03T14:04:01.214851Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest:  0.9491525423728814\n",
      "[[210   1]\n",
      " [ 11  14]] \n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Fitting Random Forest Classification to the Training set\n",
    "\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf_classifier = RandomForestClassifier(n_estimators=25, criterion='entropy')\n",
    "rf_classifier.fit(X_train, y_train)\n",
    "\n",
    "# Predicting the Test set results\n",
    "y_pred = rf_classifier.predict(X_test)\n",
    "\n",
    "# Making the Confusion Matrix\n",
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "cm5 = confusion_matrix(y_test, y_pred)\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "ac5 = accuracy_score(y_test, y_pred)\n",
    "print(\"Random Forest: \", ac5)\n",
    "print(cm5, \"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5452abf9",
   "metadata": {},
   "source": [
    "## Deep learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1aa2315a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:03.261545Z",
     "start_time": "2021-05-03T14:04:01.276914Z"
    }
   },
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from tensorflow.keras.layers import Dropout\n",
    "from tensorflow.keras.optimizers import SGD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c040cbe4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:03.276658Z",
     "start_time": "2021-05-03T14:04:03.262505Z"
    }
   },
   "outputs": [],
   "source": [
    "early_stop = EarlyStopping(monitor='val_loss',mode='min',verbose=1,patience=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "e0a9a4da",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:05.009023Z",
     "start_time": "2021-05-03T14:04:03.277641Z"
    }
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Dense(26, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "model.add(Dense(30, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "#50\n",
    "model.add(Dense(50, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "model.add(Dense(100,activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "model.add(Dense(200,activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "model.add(Dense(50,activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "#Binary Classification\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "# model.compile(optimizer='rmsprop',loss='mse')\n",
    "model.compile(optimizer= 'adam', loss='binary_crossentropy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "a6ad89fb",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.383904Z",
     "start_time": "2021-05-03T14:04:05.010023Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/4000\n",
      "29/29 [==============================] - 0s 10ms/step - loss: 0.4741 - val_loss: 0.3402\n",
      "Epoch 2/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.4298 - val_loss: 0.3539\n",
      "Epoch 3/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.4214 - val_loss: 0.3574\n",
      "Epoch 4/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.4064 - val_loss: 0.3674\n",
      "Epoch 5/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3987 - val_loss: 0.3342\n",
      "Epoch 6/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3944 - val_loss: 0.3520\n",
      "Epoch 7/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3659 - val_loss: 0.3196\n",
      "Epoch 8/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3688 - val_loss: 0.3567\n",
      "Epoch 9/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3702 - val_loss: 0.3260\n",
      "Epoch 10/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3836 - val_loss: 0.3382\n",
      "Epoch 11/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3583 - val_loss: 0.3292\n",
      "Epoch 12/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3528 - val_loss: 0.3277\n",
      "Epoch 13/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3394 - val_loss: 0.3220\n",
      "Epoch 14/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3589 - val_loss: 0.3266\n",
      "Epoch 15/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3412 - val_loss: 0.3341\n",
      "Epoch 16/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3574 - val_loss: 0.3211\n",
      "Epoch 17/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3473 - val_loss: 0.3121\n",
      "Epoch 18/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3386 - val_loss: 0.3211\n",
      "Epoch 19/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3487 - val_loss: 0.3266\n",
      "Epoch 20/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3415 - val_loss: 0.3196\n",
      "Epoch 21/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3282 - val_loss: 0.3205\n",
      "Epoch 22/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3298 - val_loss: 0.3258\n",
      "Epoch 23/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3195 - val_loss: 0.3158\n",
      "Epoch 24/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3342 - val_loss: 0.3081\n",
      "Epoch 25/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3174 - val_loss: 0.3181\n",
      "Epoch 26/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3177 - val_loss: 0.3311\n",
      "Epoch 27/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3141 - val_loss: 0.3275\n",
      "Epoch 28/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3028 - val_loss: 0.3358\n",
      "Epoch 29/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3148 - val_loss: 0.3504\n",
      "Epoch 30/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3097 - val_loss: 0.3336\n",
      "Epoch 31/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3028 - val_loss: 0.3231\n",
      "Epoch 32/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3017 - val_loss: 0.3227\n",
      "Epoch 33/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2929 - val_loss: 0.3207\n",
      "Epoch 34/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2750 - val_loss: 0.3157\n",
      "Epoch 35/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.3201 - val_loss: 0.3009\n",
      "Epoch 36/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2962 - val_loss: 0.3038\n",
      "Epoch 37/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2977 - val_loss: 0.3222\n",
      "Epoch 38/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2821 - val_loss: 0.3368\n",
      "Epoch 39/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2954 - val_loss: 0.3223\n",
      "Epoch 40/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2771 - val_loss: 0.3498\n",
      "Epoch 41/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2898 - val_loss: 0.3112\n",
      "Epoch 42/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2681 - val_loss: 0.3248\n",
      "Epoch 43/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2783 - val_loss: 0.3014\n",
      "Epoch 44/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2840 - val_loss: 0.3530\n",
      "Epoch 45/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2591 - val_loss: 0.3073\n",
      "Epoch 46/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2770 - val_loss: 0.3136\n",
      "Epoch 47/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2764 - val_loss: 0.3079\n",
      "Epoch 48/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2706 - val_loss: 0.3451\n",
      "Epoch 49/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2581 - val_loss: 0.3195\n",
      "Epoch 50/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2637 - val_loss: 0.3069\n",
      "Epoch 51/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2763 - val_loss: 0.3096\n",
      "Epoch 52/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2726 - val_loss: 0.2900\n",
      "Epoch 53/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2563 - val_loss: 0.3032\n",
      "Epoch 54/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2726 - val_loss: 0.3274\n",
      "Epoch 55/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2623 - val_loss: 0.3049\n",
      "Epoch 56/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2797 - val_loss: 0.3121\n",
      "Epoch 57/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2714 - val_loss: 0.3011\n",
      "Epoch 58/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2506 - val_loss: 0.3145\n",
      "Epoch 59/4000\n",
      "29/29 [==============================] - 0s 6ms/step - loss: 0.2556 - val_loss: 0.3097\n",
      "Epoch 60/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2579 - val_loss: 0.3059\n",
      "Epoch 61/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2516 - val_loss: 0.3205\n",
      "Epoch 62/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2694 - val_loss: 0.3104\n",
      "Epoch 63/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2556 - val_loss: 0.3036\n",
      "Epoch 64/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2521 - val_loss: 0.3031\n",
      "Epoch 65/4000\n",
      "29/29 [==============================] - 0s 6ms/step - loss: 0.2533 - val_loss: 0.2910\n",
      "Epoch 66/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2299 - val_loss: 0.3026\n",
      "Epoch 67/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2516 - val_loss: 0.3169\n",
      "Epoch 68/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2496 - val_loss: 0.3154\n",
      "Epoch 69/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2475 - val_loss: 0.2845\n",
      "Epoch 70/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2428 - val_loss: 0.3161\n",
      "Epoch 71/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2492 - val_loss: 0.2971\n",
      "Epoch 72/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2270 - val_loss: 0.2942\n",
      "Epoch 73/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2570 - val_loss: 0.3060\n",
      "Epoch 74/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2455 - val_loss: 0.3099\n",
      "Epoch 75/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2490 - val_loss: 0.3133\n",
      "Epoch 76/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2371 - val_loss: 0.3193\n",
      "Epoch 77/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2404 - val_loss: 0.3319\n",
      "Epoch 78/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2670 - val_loss: 0.3042\n",
      "Epoch 79/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2510 - val_loss: 0.2867\n",
      "Epoch 80/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2481 - val_loss: 0.3048\n",
      "Epoch 81/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2495 - val_loss: 0.3040\n",
      "Epoch 82/4000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2363 - val_loss: 0.2894\n",
      "Epoch 83/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2468 - val_loss: 0.2757\n",
      "Epoch 84/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2318 - val_loss: 0.2685\n",
      "Epoch 85/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2503 - val_loss: 0.3065\n",
      "Epoch 86/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2516 - val_loss: 0.2995\n",
      "Epoch 87/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2442 - val_loss: 0.3110\n",
      "Epoch 88/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2274 - val_loss: 0.2879\n",
      "Epoch 89/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2449 - val_loss: 0.3163\n",
      "Epoch 90/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2464 - val_loss: 0.2964\n",
      "Epoch 91/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2452 - val_loss: 0.3042\n",
      "Epoch 92/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2241 - val_loss: 0.2909\n",
      "Epoch 93/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2337 - val_loss: 0.3033\n",
      "Epoch 94/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2169 - val_loss: 0.2892\n",
      "Epoch 95/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2388 - val_loss: 0.2996\n",
      "Epoch 96/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2320 - val_loss: 0.2934\n",
      "Epoch 97/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2344 - val_loss: 0.3176\n",
      "Epoch 98/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2234 - val_loss: 0.2970\n",
      "Epoch 99/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2319 - val_loss: 0.2943\n",
      "Epoch 100/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2148 - val_loss: 0.3060\n",
      "Epoch 101/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2205 - val_loss: 0.2822\n",
      "Epoch 102/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2459 - val_loss: 0.2895\n",
      "Epoch 103/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2222 - val_loss: 0.2913\n",
      "Epoch 104/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2221 - val_loss: 0.2845\n",
      "Epoch 105/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2192 - val_loss: 0.2979\n",
      "Epoch 106/4000\n",
      "29/29 [==============================] - 0s 5ms/step - loss: 0.2320 - val_loss: 0.2926\n",
      "Epoch 107/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2180 - val_loss: 0.2892\n",
      "Epoch 108/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2388 - val_loss: 0.2944\n",
      "Epoch 109/4000\n",
      "29/29 [==============================] - 0s 4ms/step - loss: 0.2411 - val_loss: 0.2923\n",
      "Epoch 00109: early stopping\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x242b16860a0>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(x=X_train,\n",
    "          y=y_train,\n",
    "          validation_data=(X_test, y_test),\n",
    "          epochs=4000,\n",
    "          callbacks=[early_stop],\n",
    "          batch_size=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "ef3e6f0b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.557049Z",
     "start_time": "2021-05-03T14:04:20.385896Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "losses = pd.DataFrame(model.history.history)\n",
    "losses.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "26945fad",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.745399Z",
     "start_time": "2021-05-03T14:04:20.558047Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-15-22d9b303c4cd>:1: Sequential.predict_classes (from tensorflow.python.keras.engine.sequential) is deprecated and will be removed after 2021-01-01.\n",
      "Instructions for updating:\n",
      "Please use instead:* `np.argmax(model.predict(x), axis=-1)`,   if your model does multi-class classification   (e.g. if it uses a `softmax` last-layer activation).* `(model.predict(x) > 0.5).astype(\"int32\")`,   if your model does binary classification   (e.g. if it uses a `sigmoid` last-layer activation).\n"
     ]
    }
   ],
   "source": [
    "prediction = model.predict_classes(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "40dec988",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.823149Z",
     "start_time": "2021-05-03T14:04:20.747350Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.49526432e-01],\n",
       "       [8.93274415e-03],\n",
       "       [4.01787907e-02],\n",
       "       [9.29949939e-01],\n",
       "       [2.39325594e-02],\n",
       "       [4.67375247e-03],\n",
       "       [4.12780568e-02],\n",
       "       [4.91704941e-02],\n",
       "       [1.16974600e-01],\n",
       "       [1.34848922e-01],\n",
       "       [1.39942050e-01],\n",
       "       [3.89035232e-02],\n",
       "       [4.32139117e-04],\n",
       "       [8.24165791e-02],\n",
       "       [7.28726164e-02],\n",
       "       [5.21961331e-01],\n",
       "       [5.52586168e-02],\n",
       "       [1.22893870e-01],\n",
       "       [4.09991965e-02],\n",
       "       [1.11273281e-01],\n",
       "       [1.02769732e-01],\n",
       "       [6.32380322e-02],\n",
       "       [1.39964432e-01],\n",
       "       [8.26786831e-02],\n",
       "       [5.97942434e-02],\n",
       "       [7.83503056e-02],\n",
       "       [7.98841491e-02],\n",
       "       [3.35047115e-03],\n",
       "       [1.04865046e-06],\n",
       "       [1.81673139e-01],\n",
       "       [1.35643646e-01],\n",
       "       [4.14656736e-02],\n",
       "       [3.88564989e-02],\n",
       "       [1.77005947e-01],\n",
       "       [2.58275479e-01],\n",
       "       [4.31751102e-01],\n",
       "       [2.28770729e-02],\n",
       "       [1.89442909e-03],\n",
       "       [8.27400684e-02],\n",
       "       [8.24692717e-04],\n",
       "       [4.99166362e-02],\n",
       "       [1.31468102e-01],\n",
       "       [3.46634388e-02],\n",
       "       [8.93195122e-02],\n",
       "       [9.20221567e-01],\n",
       "       [6.28971756e-02],\n",
       "       [1.04788557e-01],\n",
       "       [1.07200980e-01],\n",
       "       [1.10820130e-01],\n",
       "       [1.41186357e-01],\n",
       "       [7.63416067e-02],\n",
       "       [2.03030761e-02],\n",
       "       [9.06885266e-02],\n",
       "       [1.36122525e-01],\n",
       "       [2.36326084e-01],\n",
       "       [5.35017401e-02],\n",
       "       [8.60878527e-02],\n",
       "       [8.83050635e-02],\n",
       "       [1.03156308e-04],\n",
       "       [9.16556954e-01],\n",
       "       [9.09704387e-01],\n",
       "       [2.40069684e-02],\n",
       "       [1.74481813e-02],\n",
       "       [8.50445256e-02],\n",
       "       [5.28743211e-03],\n",
       "       [8.50686058e-02],\n",
       "       [8.20974354e-03],\n",
       "       [9.01061147e-02],\n",
       "       [8.68368596e-02],\n",
       "       [5.98295592e-03],\n",
       "       [6.85149245e-03],\n",
       "       [9.01256353e-02],\n",
       "       [9.82489511e-02],\n",
       "       [1.13083281e-01],\n",
       "       [1.03026055e-01],\n",
       "       [1.05408303e-01],\n",
       "       [5.07874275e-03],\n",
       "       [1.52343720e-01],\n",
       "       [1.77127158e-03],\n",
       "       [2.25522872e-02],\n",
       "       [1.07419349e-01],\n",
       "       [8.60302627e-01],\n",
       "       [7.79684633e-02],\n",
       "       [3.79808359e-02],\n",
       "       [1.02770559e-01],\n",
       "       [1.35664165e-01],\n",
       "       [9.35059726e-01],\n",
       "       [6.86107874e-02],\n",
       "       [8.93730819e-02],\n",
       "       [1.85007811e-01],\n",
       "       [1.14096671e-01],\n",
       "       [3.10648838e-03],\n",
       "       [1.72957003e-01],\n",
       "       [8.35713297e-02],\n",
       "       [8.89033601e-02],\n",
       "       [1.34401143e-01],\n",
       "       [1.27226606e-01],\n",
       "       [9.07686114e-01],\n",
       "       [2.78102700e-02],\n",
       "       [1.31994963e-01],\n",
       "       [2.07979162e-03],\n",
       "       [1.12929448e-01],\n",
       "       [1.01360127e-01],\n",
       "       [8.28668028e-02],\n",
       "       [1.05932184e-01],\n",
       "       [9.64242741e-02],\n",
       "       [4.23488170e-02],\n",
       "       [2.99787428e-03],\n",
       "       [1.11907050e-01],\n",
       "       [1.35107666e-01],\n",
       "       [8.63682508e-01],\n",
       "       [9.20282423e-01],\n",
       "       [8.64548832e-02],\n",
       "       [1.68005019e-01],\n",
       "       [1.73215370e-03],\n",
       "       [4.63520959e-02],\n",
       "       [1.31814498e-02],\n",
       "       [5.28320447e-02],\n",
       "       [9.16392982e-01],\n",
       "       [1.75056279e-01],\n",
       "       [3.83402407e-03],\n",
       "       [9.99539569e-02],\n",
       "       [9.61222202e-02],\n",
       "       [8.31144601e-02],\n",
       "       [1.40614331e-01],\n",
       "       [1.07129142e-01],\n",
       "       [1.06244229e-01],\n",
       "       [1.07183360e-01],\n",
       "       [8.24984387e-02],\n",
       "       [2.60149222e-02],\n",
       "       [1.39597207e-01],\n",
       "       [1.00484021e-01],\n",
       "       [1.18641324e-01],\n",
       "       [1.83533728e-02],\n",
       "       [1.16974324e-01],\n",
       "       [3.81344445e-02],\n",
       "       [1.05113186e-01],\n",
       "       [1.57682803e-02],\n",
       "       [9.26913396e-02],\n",
       "       [8.11034888e-02],\n",
       "       [1.07196875e-01],\n",
       "       [1.66182965e-01],\n",
       "       [1.11016929e-01],\n",
       "       [1.95094775e-02],\n",
       "       [1.79436784e-02],\n",
       "       [3.39814164e-02],\n",
       "       [1.68260589e-01],\n",
       "       [1.51988611e-01],\n",
       "       [9.28170830e-02],\n",
       "       [8.27375799e-02],\n",
       "       [1.12174809e-01],\n",
       "       [8.25431272e-02],\n",
       "       [9.54386145e-02],\n",
       "       [1.09280303e-01],\n",
       "       [2.76433257e-03],\n",
       "       [9.14574683e-01],\n",
       "       [1.13444313e-01],\n",
       "       [6.78974614e-02],\n",
       "       [2.89686974e-02],\n",
       "       [1.75978765e-01],\n",
       "       [1.08296655e-01],\n",
       "       [8.29594433e-02],\n",
       "       [5.85323723e-04],\n",
       "       [3.61566655e-02],\n",
       "       [1.94121543e-02],\n",
       "       [3.93179394e-02],\n",
       "       [1.18425839e-01],\n",
       "       [9.20350075e-01],\n",
       "       [4.12789136e-02],\n",
       "       [1.06758140e-01],\n",
       "       [1.96093637e-02],\n",
       "       [5.46499109e-03],\n",
       "       [1.50395885e-01],\n",
       "       [1.77841067e-01],\n",
       "       [2.22856309e-02],\n",
       "       [7.82846436e-02],\n",
       "       [3.51220705e-02],\n",
       "       [1.53500751e-01],\n",
       "       [4.79809456e-02],\n",
       "       [7.20962882e-02],\n",
       "       [5.85429966e-02],\n",
       "       [3.46042573e-01],\n",
       "       [1.64918713e-02],\n",
       "       [4.64999042e-02],\n",
       "       [1.05761457e-02],\n",
       "       [4.69026947e-03],\n",
       "       [8.62548649e-02],\n",
       "       [9.38913703e-01],\n",
       "       [8.94228399e-01],\n",
       "       [4.11801822e-02],\n",
       "       [7.26155862e-02],\n",
       "       [4.76261154e-02],\n",
       "       [1.07651889e-01],\n",
       "       [4.23631221e-01],\n",
       "       [5.12374826e-02],\n",
       "       [9.32970122e-02],\n",
       "       [9.21949089e-01],\n",
       "       [9.61846560e-02],\n",
       "       [5.09677976e-02],\n",
       "       [3.66175771e-02],\n",
       "       [1.06506445e-01],\n",
       "       [8.76481365e-03],\n",
       "       [1.07197016e-01],\n",
       "       [9.28507566e-01],\n",
       "       [9.28600669e-01],\n",
       "       [9.96473245e-05],\n",
       "       [2.46375985e-02],\n",
       "       [3.89215425e-02],\n",
       "       [9.17017281e-01],\n",
       "       [1.43109800e-08],\n",
       "       [8.27376693e-02],\n",
       "       [3.99140990e-04],\n",
       "       [4.47335355e-02],\n",
       "       [8.17579180e-02],\n",
       "       [1.38770204e-07],\n",
       "       [6.69010263e-03],\n",
       "       [8.27004761e-02],\n",
       "       [9.56344530e-02],\n",
       "       [4.75475676e-02],\n",
       "       [8.53046104e-02],\n",
       "       [9.01445448e-02],\n",
       "       [2.32007712e-01],\n",
       "       [8.27035680e-02],\n",
       "       [1.98457390e-03],\n",
       "       [1.78881779e-01],\n",
       "       [2.00347306e-04],\n",
       "       [9.14629176e-03],\n",
       "       [9.20353413e-01],\n",
       "       [9.37377140e-02],\n",
       "       [4.66021337e-02],\n",
       "       [8.46213847e-02],\n",
       "       [8.70716583e-04],\n",
       "       [1.35704309e-01],\n",
       "       [9.47060883e-02],\n",
       "       [9.61436480e-02],\n",
       "       [5.77432215e-02]], dtype=float32)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "70cb4ea0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.838623Z",
     "start_time": "2021-05-03T14:04:20.824145Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report,confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "cc2cfd11",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.854580Z",
     "start_time": "2021-05-03T14:04:20.839621Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.97      0.96       211\n",
      "           1       0.65      0.52      0.58        25\n",
      "\n",
      "    accuracy                           0.92       236\n",
      "   macro avg       0.80      0.74      0.77       236\n",
      "weighted avg       0.91      0.92      0.92       236\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test,prediction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "16237291",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.870538Z",
     "start_time": "2021-05-03T14:04:20.856577Z"
    }
   },
   "outputs": [],
   "source": [
    "cm6 = confusion_matrix(y_test,prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "74aeb32a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:20.885497Z",
     "start_time": "2021-05-03T14:04:20.871536Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Deep Learning accuracy :  0.9491525423728814\n",
      "[[204   7]\n",
      " [ 12  13]]\n"
     ]
    }
   ],
   "source": [
    "ac6 = accuracy_score(y_test, y_pred)\n",
    "print(\"Deep Learning accuracy : \",ac6)\n",
    "print(cm6)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63e5b361",
   "metadata": {
    "variables": {
     "cm1[0][0]": "209",
     "cm1[1][1]": "7",
     "cm2[0][0]": "201",
     "cm2[1][1]": "14",
     "cm3[0][0]": "211",
     "cm3[1][1]": "6",
     "cm4[0][0]": "199",
     "cm4[1][1]": "15",
     "cm5[0][0]": "210",
     "cm5[1][1]": "14",
     "cm6[0][0]": "204",
     "cm6[1][1]": "13",
     "int(ac1*100)": "91",
     "int(ac2*100)": "91",
     "int(ac3*100)": "91",
     "int(ac4*100)": "90",
     "int(ac5*100)": "94",
     "int(ac6*100)": "94"
    }
   },
   "source": [
    "## MODEL\n",
    "\n",
    "\n",
    "| Model | Name| Acuuracy(%)|True|False|\n",
    "| :---        |    :----:   |   :---: | :---:|    :----:   \n",
    "|Logestic Regression| LR_classifier| {{int(ac1*100)}}|{{cm1[0][0]}}|{{cm1[1][1]}}|\n",
    "|K-NN |knn_classifier | {{int(ac2*100)}}| {{cm2[0][0]}} |{{cm2[1][1]}} |\n",
    "| SVM |svm_rbf_classifier | {{int(ac3*100)}} | {{cm3[0][0]}} | {{cm3[1][1]}} |\n",
    "|Desision Tree  | dt_classifier| {{int(ac4*100)}} |{{cm4[0][0]}} | {{cm4[1][1]}} |\n",
    "| random forest | rf_classifier | {{int(ac5*100)}}|{{cm5[0][0]}}|{{cm5[1][1]}}|\n",
    "| Deep Learning | model | {{int(ac6*100)}}|{{cm6[0][0]}}|{{cm6[1][1]}}|\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "b6503076",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.198209Z",
     "start_time": "2021-05-03T14:04:21.168255Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>model</th>\n",
       "      <th>acc</th>\n",
       "      <th>True</th>\n",
       "      <th>False</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Logestic Regression</td>\n",
       "      <td>91.525424</td>\n",
       "      <td>209</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>K-NN</td>\n",
       "      <td>91.101695</td>\n",
       "      <td>201</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>SVM</td>\n",
       "      <td>91.949153</td>\n",
       "      <td>211</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Desision Tree</td>\n",
       "      <td>90.677966</td>\n",
       "      <td>199</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>random forest</td>\n",
       "      <td>94.915254</td>\n",
       "      <td>210</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Deep Learning</td>\n",
       "      <td>94.915254</td>\n",
       "      <td>204</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 model        acc  True  False\n",
       "0  Logestic Regression  91.525424   209      7\n",
       "1                 K-NN  91.101695   201     14\n",
       "2                  SVM  91.949153   211      6\n",
       "3        Desision Tree  90.677966   199     15\n",
       "4        random forest  94.915254   210     14\n",
       "5        Deep Learning  94.915254   204     13"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# df = pd.DataFrame({'lab':['A', 'B', 'C'], 'val':[10, 30, 20]})\n",
    "df_models = pd.DataFrame({\n",
    "    'model': [\n",
    "        \"Logestic Regression\", \"K-NN\", \"SVM\", \"Desision Tree\", \"random forest\",\n",
    "        \"Deep Learning\"\n",
    "    ],\n",
    "    \"acc\":[ac1*100,ac2*100,ac3*100,ac4*100,ac5*100,ac6*100,],\n",
    "    \"True\":[cm1[0][0],cm2[0][0],cm3[0][0],cm4[0][0],cm5[0][0],cm6[0][0]],\n",
    "    \"False\":[cm1[1][1],cm2[1][1],cm3[1][1],cm4[1][1],cm5[1][1],cm6[1][1]]\n",
    "})\n",
    "df_models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "53d0004e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.402630Z",
     "start_time": "2021-05-03T14:04:21.200171Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='model'>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1440x1440 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAE3CAYAAABGupFbAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Z1A+gAAAACXBIWXMAAAsTAAALEwEAmpwYAAAvtElEQVR4nO3deZhcZZXH8e8vCwkkLAESRAJ2ZBllDRAQjCKaYXUQZEeQgGw6oCLoDCKyCI4oCiOCSBAEZBdkERBZDLvIEhAEBFGIhIkQCLuEJTnzx3krKZpO0ntV9/19nqeerrp169Z7u6rOfe+7nKuIwMzMqmFAowtgZma9x0HfzKxCHPTNzCrEQd/MrEIc9M3MKsRB38ysQgY1ugAAyy67bLS0tDS6GGZmfcp99933fESM7MhrmiLot7S0cO+99za6GGZmfYqkqR19jZt3zMwqxEHfzKxCHPTNzCrEQd/MrEIc9M3MKsRB38ysQhz0zcwqxEHfzKxCmmJylpn1LWuds1avvt9DEx/q1ffrz1zTNzOrEAd9M7MKcdA3M6sQB30zswpxR66ZWSv9uaPaNX0zswpxTd8apj/XpsyaVZ8O+g4aZmYds9CgL2kocCswpKx/aUQcJWkMcBGwDHAf8PmIeEvSEOBcYH3gBWCXiHiqh8rfr/mgZmbdrT01/TeBT0XEa5IGA7dL+i1wCHBSRFwk6WfAPsBp5e+LEbGKpF2B7wO79FD5zZpWbx60e/uA/dCT/+jV97Pus9CgHxEBvFYeDi63AD4FfK4sPwc4mgz625b7AJcCp0hS2Y7ZXA4cZr2vXW36kgaSTTirAKcCfwNeioh3yirTgBXK/RWApwEi4h1JL5NNQM93Y7kBBw0zs45qV9CPiNnAWElLAZcDH+rqG0vaH9gfYKWVVurq5szMuk1/rlB2aJx+RLwETAY2BpaSVDtojAaeKfefAVYEKM8vSXbott7WpIgYFxHjRo4c2bnSm5lZhyw06EsaWWr4SFoU2Ax4lAz+O5bVJgJXlvtXlceU53/v9nwzs+bQnuad5YFzSrv+AOCSiLha0iPARZKOA+4Hzizrnwn8UtITwExg1x4ot5mZdUJ7Ru88CKzbxvK/Axu2sXwWsFO3lM7MzLpVn56R29/1584kM2sMB32zHuKDtjUjZ9k0M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0Dczq5CFBn1JK0qaLOkRSQ9L+mpZfrSkZyQ9UG5b173mm5KekPSYpC16cgfMzKz9BrVjnXeAQyNiiqTFgfsk3VCeOykifli/sqTVgV2BNYD3AzdKWi0iZndnwc3MrOMWWtOPiOkRMaXcfxV4FFhhAS/ZFrgoIt6MiCeBJ4ANu6OwZmbWNR1q05fUAqwL/LEsOkjSg5LOkjSiLFsBeLruZdNY8EHCzMx6SbuDvqThwGXAwRHxCnAasDIwFpgO/Kgjbyxpf0n3Srp3xowZHXmpmZl1UruCvqTBZMA/PyJ+DRARz0bE7IiYA5zBvCacZ4AV614+uix7l4iYFBHjImLcyJEju7IPZmbWTu0ZvSPgTODRiDixbvnydat9FvhzuX8VsKukIZLGAKsCd3dfkc3MrLPaM3pnPPB54CFJD5RlhwO7SRoLBPAUcABARDws6RLgEXLkz4EeuWNm1hwWGvQj4nZAbTx17QJe813gu10ol5mZ9QDPyDUzqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpZaNCXtKKkyZIekfSwpK+W5UtLukHSX8vfEWW5JJ0s6QlJD0par6d3wszM2qc9Nf13gEMjYnVgI+BASasDhwE3RcSqwE3lMcBWwKrltj9wWreX2szMOmWhQT8ipkfElHL/VeBRYAVgW+Ccsto5wHbl/rbAuZHuApaStHx3F9zMzDquQ236klqAdYE/AstFxPTy1D+B5cr9FYCn6142rSwzM7MGa3fQlzQcuAw4OCJeqX8uIgKIjryxpP0l3Svp3hkzZnTkpWZm1kntCvqSBpMB//yI+HVZ/Gyt2ab8fa4sfwZYse7lo8uyd4mISRExLiLGjRw5srPlNzOzDmjP6B0BZwKPRsSJdU9dBUws9ycCV9Yt37OM4tkIeLmuGcjMzBpoUDvWGQ98HnhI0gNl2eHA8cAlkvYBpgI7l+euBbYGngD+BezdnQU2M7POW2jQj4jbAc3n6QltrB/AgV0sl5mZ9QDPyDUzqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6sQB30zswpx0DczqxAHfTOzCnHQNzOrEAd9M7MKcdA3M6uQhQZ9SWdJek7Sn+uWHS3pGUkPlNvWdc99U9ITkh6TtEVPFdzMzDquPTX9s4Et21h+UkSMLbdrASStDuwKrFFe81NJA7ursGZm1jULDfoRcSsws53b2xa4KCLejIgngSeADbtQPjMz60ZdadM/SNKDpflnRFm2AvB03TrTyjIzM2sCnQ36pwErA2OB6cCPOroBSftLulfSvTNmzOhkMczMrCM6FfQj4tmImB0Rc4AzmNeE8wywYt2qo8uytrYxKSLGRcS4kSNHdqYYZmbWQZ0K+pKWr3v4WaA2sucqYFdJQySNAVYF7u5aEc3MrLsMWtgKki4ENgWWlTQNOArYVNJYIICngAMAIuJhSZcAjwDvAAdGxOweKbmZmXXYQoN+ROzWxuIzF7D+d4HvdqVQZmbWMzwj18ysQhz0zcwqxEHfzKxCHPTNzCrEQd/MrEIc9M3MKsRB38ysQhz0zcwqxEHfzKxCHPTNzCrEQd/MrEIc9M3MKsRB38ysQhz0zcwqxEHfzKxCHPTNzCrEQd/MrEIc9M3MKsRB38ysQhz0zcwqxEHfzKxCHPTNzCrEQd/MrEIc9M3MKmShQV/SWZKek/TnumVLS7pB0l/L3xFluSSdLOkJSQ9KWq8nC29mZh3Tnpr+2cCWrZYdBtwUEasCN5XHAFsBq5bb/sBp3VNMMzPrDgsN+hFxKzCz1eJtgXPK/XOA7eqWnxvpLmApSct3U1nNzKyLOtumv1xETC/3/wksV+6vADxdt960sszMzJpAlztyIyKA6OjrJO0v6V5J986YMaOrxTAzs3bobNB/ttZsU/4+V5Y/A6xYt97osuw9ImJSRIyLiHEjR47sZDHMzKwjOhv0rwImlvsTgSvrlu9ZRvFsBLxc1wxkZmYNNmhhK0i6ENgUWFbSNOAo4HjgEkn7AFOBncvq1wJbA08A/wL27oEym1mDtcy6oFff76lefbf+vX8LDfoRsdt8nprQxroBHNjVQpmZWc9YaNBvZv35aAzev+72VK++m1lz6tNB36yZ9eZB7aleeyfr65x7x8ysQhz0zcwqxEHfzKxCHPTNzCrEQd/MrEKadvTO22+/zbRp05g1a9Z81znjM72bwPPRRx/tsW0PHTqU0aNHM3jw4B57DzOzpg3606ZNY/HFF6elpQVJba7z9rSXerVMHx69VI9sNyJ44YUXmDZtGmPGjOmR9zAzgyZu3pk1axbLLLPMfAN+fyKJZZZZZoFnNWZm3aFpgz5QiYBfU6V9NbPGaeqgb2Zm3atp2/Rbaznsmm7d3lUHje/W7ZmZ9QWu6S/Ewfvszq5bb8pnJ2zMpEmTALjuuutYb731WGeddZgwIZONvvbaa+y9996stdZarL322lx22WWNLLaZWZv6TE2/UY754SksOWIEs954gy98djO23XZb9ttvP2699VbGjBnDzJl5zfhjjz2WJZdckoceegiAF198sZHFNjNrk4P+Qlzwi9P5/XVXA/DPZ55m0qRJbLLJJnOHVi699NIA3HjjjVx00UVzXzdixIjeL6yZ2UK4eWcB7vnD7dx1+82ce+X1/Or621l33XUZO3Zso4tlZtZpDvoL8Norr7DEkkux6KKL8eQTj3PXXXcxa9Ysbr31Vp588kmAuc07m222Gaeeeurc17p5x8yakYP+AozfdAKz33mH7T75EX78vWPYaKONGDlyJJMmTWL77bdnnXXWYZdddgHgiCOO4MUXX2TNNddknXXWYfLkyQ0uvZnZe/WZNv2njv/0e5Y92MNpGBYZMoSf/vLSuY/XrkvDsNVWW71r3eHDh3POOef0aHnMzLrKNX0zswpx0DczqxAHfTOzCulSm76kp4BXgdnAOxExTtLSwMVAC/AUsHNEeCiLmVkT6I6a/icjYmxEjCuPDwNuiohVgZvKYzMzawI90byzLVAbxnIOsF0PvIeZmXVCV4dsBnC9pABOj4hJwHIRMb08/09guS6+R0O89OJM9t91WwCen/EcAwYM5P3vGwXA3XffzSKLLNLI4pmZdUpXg/7HIuIZSaOAGyT9pf7JiIhyQHgPSfsD+wOstNJKC3+no5d8z6K1O1Hgmgf3nbrA55casTSX/O42AE478XgWW2wYPzru23Off+eddxg0qM9MczAzA7oY9CPimfL3OUmXAxsCz0paPiKmS1oeeG4+r50ETAIYN25cmweGZrPXXnsxdOhQ7r//fsaPH88SSyzB8OHD+frXvw7AmmuuydVXX01LSwvnnXceJ598Mm+99RYf+chH+OlPf8rAgQMbvAdmVnWdbtOXNEzS4rX7wObAn4GrgIlltYnAlV0tZDOZNm0ad955JyeeeOJ813n00Ue5+OKLueOOO3jggQcYOHAg559/fi+W0sysbV2p6S8HXF6u7ToIuCAirpN0D3CJpH2AqcDOXS9m89hpp50WWmO/6aabuO+++9hggw0AeOONNxg1alRvFM/MbIE6HfQj4u/AOm0sfwGY0JVCNbNhw4bNvT9o0CDmzJkz9/GsWbMAiAgmTpzI9773vV4vn5nZgnhGbhe0tLQwZcoUAKZMmTI33fKECRO49NJLee657M6YOXMmU6cuuOPYzKw3OOh3wQ477MDMmTNZY401OOWUU1httdUAWH311TnuuOPYfPPNWXvttdlss82YPn36QrZmZtbz+s6Yw6Nffs+ink6tXPOlQ3JScX1qZYBFF12U66+/vs3X7LLLLnNz7ZuZNQvX9M3MKsRB38ysQhz0zcwqxEHfzKxCHPTNzCrEQd/MrEL6zpDNBlj3A8uw6odWn/v4umt+Q0tLS5vrDh8+nNdee62XSmZm1jl9Juivdc5a3bq98yfcttB1hgxddG56ZYCWVuP0zcz6GjfvdMBrr73GhAkTWG+99VhrrbW48sr3JhCdPn06m2yyCWPHjmXNNdfkttvyoHH99dez8cYbs95667HTTjv5rMDMGsJBfwHenPUGO2/xcXbe4uMcvO8eDB06lMsvv5wpU6YwefJkDj30UCLefSmACy64gC222IIHHniAP/3pT4wdO5bnn3+e4447jhtvvJEpU6Ywbty4BaZmNjPrKX2meacRWjfvRASHH344t956KwMGDOCZZ57h2Wef5X3ve9/cdTbYYAO+8IUv8Pbbb7PddtsxduxYbrnlFh555BHGjx8PwFtvvcXGG2/c6/tjZuag3wHnn38+M2bM4L777mPw4MG0tLTMTadcs8kmm3DrrbdyzTXXsNdee3HIIYcwYsQINttsMy688MIGldzMLLl5pwNefvllRo0axeDBg5k8eXKb6ZKnTp3Kcsstx3777ce+++7LlClT2Gijjbjjjjt44oknAHj99dd5/PHHe7v4Zmau6XfE7rvvzjbbbMNaa63FuHHj+NCHPvSedW6++WZOOOEEBg8ezPDhwzn33HMZOXIkZ599NrvtthtvvvkmAMcdd9zcVMxmZr2lzwT9hyY+9J5lPZ1a+a7Hpr3r8bLLLssf/vCHNtetjcaZOHEiEydOfM/zn/rUp7jnnnu6v5BmZh3g5h0zswpx0DczqxAHfTOzCmnqoN964lN/VqV9NbPGadqgP3ToUF544YVKBMOI4IUXXmDo0KGNLoqZ9XNNO3pn9OjRTJs2jRkzZsx3nWdffKMXSwSPvrpoj2176NChjB49use2b2YGPRj0JW0J/BgYCPw8Io7vyOsHDx7MmDFjFrjOVodd0/kCdsJTx3+6V9/PzKy79UjzjqSBwKnAVsDqwG6SVl/wq8zMrKf1VJv+hsATEfH3iHgLuAjYtofey8zM2qmngv4KwNN1j6eVZWZm1kDqidExknYEtoyIfcvjzwMfiYiD6tbZH9i/PPw34LFuL8j8LQs834vv19u8f31Xf9438P51tw9ExMiOvKCnOnKfAVasezy6LJsrIiYBk3ro/RdI0r0RMa4R790bvH99V3/eN/D+NYOeat65B1hV0hhJiwC7Alf10HuZmVk79UhNPyLekXQQ8DtyyOZZEfFwT7yXmZm1X4+N04+Ia4Fre2r7XdSQZqVe5P3ru/rzvoH3r+F6pCPXzMyaU9Pm3jHrKElqdBnMmp2DvvVpkpaU9E1JQ8KnrVZIGtzoMvS0zlZyHPSbhKQe+yyUlpX0I0k9lzWuMb4DfBf4MvTs/7GvK+lR6h/3uzMjSUMlbQSsUh4v3eAi9QhJAzpbyfEPpAlIUkTMKfc/Immd7tx+pOeBzYDdu3PbjSBpRN3DXwOTgb0lrRIRc/pjMOsOETEbQNKG5XF/PDN6C/gQ8DVJNwL7Seo3Octr3+3yPV9U0pGSPtihbfTPz73vkbQCWWNdC1gUOBi4oSs/TEnfAF4HLo+I6ZK2ImvG20TEP7te6t4laQhwOfn/2S4iXpb0SWADYAiwTkTs2MgyNjNJHwOOA14C7gcejohLG1qobiBpYO2AVh7vDpwDXBQRezSuZD1H0qFkMsv/AO6OiG3a+1rX9BugjdPsDYBTgOciYn3gJGAvMj1FZ7a/VJkUtwXwfeA8SR+MiN8CfwIO6ULxe11pnjqmPHwL2Ag4QNLWZPDaGTgXWK4sq3wzTxvfscWAzwJfAr4GbA2M7w9t33VnMPtKOpWcH/QN4DFJa5Tn+uTZX/nuD2i1bEtywut3yM9xI0mbt3eblf5h9La6U7Pal7R21ZRngLeB95fnzwBmA5tJGt7ebUsaWPIenVSymx4PnAUIOEjSUcBRwJaS1uq+Petxg4H1gY8C3yRzmzwOfJWs7fweWIkcI/0NyNPfhpS0SdR9xz5cFg0mz4g+B1wMXB0RXwPeaUwJO69811X3eISk3wJbAheWpsyrgBHAv0M2ZbU+EDa7Wrt9acoZWbfP7yc/v6kRcR/wbeAH7d2ug34vKO30i9SaaiRtLOke4FRJ1wP/Imuqr0r6eHnZmWQ66oW270saVL4cs4FngdmStgVuB54ia/cnApsCOwFBtoE37eff6mA3GHgFGBoRjwK3AesBx5LNYNsDc4ALgIGSDuvd0jaH+s9T0kRJfwQOl/QVYHngYWBzYNOIOLasurWkxXu/tJ1TmnKiVRBfDXinNO09KWk1YDpwI7CypP0kHQms2qBit1upuC0Bc9vtB0r6LnAv8H1J6wL/B2xc97Jfkmlv9irbWODvuml/9P3MJ4FRAGX0zH8DJ0bEtsAM4FDg78BU8kc4ICJuBm4FXlzQhiX9B3CfpAPKojuBu8hT+UXIALk4+YXfhqwhL828YNlUJK0gaRfgB7UOuIh4nfwRf76s9nVgD/J/cwjZzj+gHPQOI/e/EiSNLQf4WpBYWtKawMeATYBrgAPIM6UbyLOkDSRtKOlm8jvR9B179WfJpaJ/NHCypC0j4o9Ai6RbyFrveeR+30T+hvYCBkbEXxpS+HZSDlA4BnhfebwlWfl7lTxjeRo4KiKuA0LS4ZJWJNv1byF/Fws/y40I33rgRn7J6h9/Dvgw2eF4MbBuWT6SDMxrA+PIC87s2IH3+ThZy70D+F+yI3hF4IfAV8o6XyKvZPa+8ng82fSzaqP/T/X/L7KT8WEySF0BnA8sX1fmC+v24Sjgyrb+11W4AcsApwF3k2eDw8gzn1PJvpwryv/zHmBiec0gcvTWz8kD456N3o927KcoA07K4+FkMD+WrEzdB3yGTGm8Ink2SPlNjSn3hzZ6Pxa2j3X3FysxYQnyQP0SsHN5bhRZq9+r7Ov3yud7ETlE9UJgpYW+X6N3uD/eWn2IK5S/p5KjCQAuAz4BLFoenwl8hRyVMgEY1ta2atsDBtc/R7Zlnw3sTQ5fPADYgWzPH1NupwKHN/p/M5//10jgSuB0YMWybHgJ+qeSHbcfBi4FhtS97mlg9/n9r/rrjey3eAH4Q3m8JDCT7NgbBHwQ+COZ6JC6dTboq/8rsjnvSPISrN8oQfFXZI1+tbr1tiaH8Z4HLN7ocrdjvwa0erxEiQ8HlcenAafV1iUPdHfWHdBGld/KT8rvY9DC3tPNOz0gIkLSR8vp5jGlje0Y4IOSxgGXAPsBO5YO1Q8Ct0fEGxFxU0S8Xnc6G7X7kpYkT9FrnbC1z+848lT+BrKj8+PAPsBSZC3hSeBmYEoP73pnLQEMj4gDIuJpSUtHxGvkj/tP5EFxNnkZzg/VvW4iOXoH6LfjzueStIyknwFrkp/v4pKWiYiXgSfJ4FfrmL0OeEc5Y3lbsua7fv33qgG70C61NunSjDNM0hfJJtE7yObJb5PNNldExKcj4nFJH5S0PPAt8gxwj4h4tVH7sDB1n8Ocsp/7SfpwRLxCHrDWlfRv5L5uIWm9yGab+8h9r43sG0J+1i9ExI51n//8NfpI1x9uvLcp59/I0+fPtFq+Hzn2HmBHcizxZGCfhWx/dPm7FfD7uuUDKEd24ATg1+X+4mTAnAn8lXK20aw3YDmydnY4ecp6MnkKfzgwtPz9GVm7/Xyjy9vA/9Ng6k7fgV8A55T76wFvkAdPyBEeJ5Cn/HcAExpd/nbs37uacsqyMcBfgHPL44Hkme0P69Y5knIW2/r1zXZrvY/k4IrLyEEXVwA7leUnA4eV+/8F3FH3mtbxZnhHyuDJWd1IeQnI68ig/2nyxxZk08RVwBNk+9vtEfF9SYtHXW1EkqLVByLp/cCjEbGkpC+TTT/Hl87e2izekRExQ9KDwBERcVVZvhHwZkTcTxMrNbuPkcPO/kaelSxPHgxejIgjJI0HfkM25/y2UWVtBmUEy2xlioF7yLb5OySdDxARu9etu2TkmUDTkvQJYEZEPFIeb0u2019FjrnfCdgO2C8iZkpaBfgtmbp9Y+AfwFcj4pk2Nt80Wv1mVyHb4S8lz8avLfFjDbJJZ1FykMJZETG5fLZfjoiZ5fXviRXtLoeDfse1/odL+nfgx2TzydeBN8nT0ZXIttXNgJfJD3E9sif+G+T/f45azShs4/1+Xra9IjA5Iq4vp4ejyM7b2RGxh6R9yNPbVRe0vWZVhrW+VRfUDiAD/7EREZKGRY7k6ffqA8R8nq/9j/4b2DwiJpRhrq+Q16O+p9cK20nKGdYnkAf4r0XENEnfIkejXEAGwMUjYndJ15KdmJdGxNtl1Mpo8kz3tgbtQodJWgr4ArAn2e92JXBJRHxH0hiyk3ZWRHxP0vfJ4dzf6WyAb1OjT3f62o33nloNI0dDrFe3bFCrddYkayUjOvmei5GTt14tX5ZVy/Ityfb8gXXr7t3o/1EX/78Dyt+VyRr/Fxtdpl7ef1HXuQd8hDy41ypo7/pb7t9b+z8BuwEtNHkzR13ZryCH325KNuWcBXywPLcYWdPfhhzgcC2wcqPL3IF9ax0rFiGD/J3AYmXZlmS/1OLl8TbknJ1NyWba+u/CgO4olztyOyiydrWEpG9IWiOy5rkY8ENJP5M0CbhfmQphZUmnkU06F0XE3DH3C5tA0eo9/0WOUR9G/jCuknQJOePwF8DcJGMR8Yvu2tfeJmkQ2YF1HtnOeV5E/KzBxepVkeYos6J+nxyhcgKwb+352l/Nm5z0Q+Az5Qz0woh4qrZesyqfNWQ/xOlkLqUnyRnWW8Lc7/3lwCci4ibyDHqlBhS3w8pnUZsVvZqkJSJnyV9CNvfWzuKuJztnjy6PbyP7tx6IiFfLd2EAdN8sczfvLEQbTTk7A0eQQ8VqP7ojyXH4fyNHF/yEnA14LTmr9oyImNXVcpBDFNcnJyXtR47rX47s3HyhK9tvFsrEc5sDF0TEm40uT29o3bxX2rh/TXbknVG+c1sDkyLizoU1/fQlko4nv8fXRMRJytxJJwHjIuLVcuD7R0ScKmnRiHijoQXugDIy7/tkv94A4D8j4klJl5L9dN8u661JjrzbJCL+2uPlctCfv1YdL5uQzSubkKNuWsgRJVPISVDvlPU2ID/oH0XENXXbWmC7fTvL81HgxxGxQVe2Y82jvlIhqSUinir3/wGcHRFHSloJ2AV4f2S+nD6lre++MnVI7TezOjlh8dMR8Q9JF5AVquHl9rWIaNbhxkCbB+7FyGaaCyPiMkkPA49HxGeVqRTOJEfq/K2sPy4i7u2Vsjrov1erH+L7yZQG25EzW88rq71GTpq4rHzAtdPUdYDvRcQVbW2vG8p2J9l++2B3btd6j6SVgf+r1VrLwfwHZBrsP5GVhnWB8yNiubp1DgR+GxHntbnhJtOq0jSYbK++IyKebWPdH5Mj0/ZVpt9Yiewnu6hXC91FJaBPixxNN5z8HE8ihyD/B/CDiDhHmQ10uWhAKnAH/Tr1R+vaF1bSr8kmlIMi4n7lsMl9ImJsWW8ImTvndLIz5qm67XV7UO6OMwZrHGUyre8AR0fES5KGkePOLyZTKnyDnKn9n2XEyp8i4puSliHTKzwREXc3qvzt0UaT6PJk2/wb5FyLSZEj0Opr+yuRk44mRsQtjSh3VyivVXAy8BA5+/mIiPizpB+Rn9lpysRpe5MDOwJYIiKm9nZZ3ZFbpy7gfxL4bhlCdTw5omKR0qHyK+B5SScrL1JyN5n349W6U/OBZXvdfkR1wO+blNkSFRGvRMTBwDhJy5KTz8YDf4y8sM0ZwJulZv9l4L+V8zBeiIgLmj3gw7zvvaRNSg1+d3LY7SfJ0TjfKuu9UzcA4R/kAW9aY0rdPpKGq9WVuJTXrtgV2D8iJpKf5w6SRpVVllZm/hxGnskNj4gXI2JqRwZ0dBcH/TrKjIV3khko9wA+W35k95OTrYaUH+YXyPH3LWTN5JDSMw84MNs8SgMiYnYZcbN8eeq75Djtf5ETdHYAiMwEuRIwqrT3folMp9BnLgKiHLn2bfKg9TqZC2pMGbHzCzKF+JfL6nNjUET8qtbG3YwkrU3W5sfVLRNZ6XsD+JIyZfqJEXFMRDxHjsZZmeyofTgitioHOKAx130YtPBV+qc2Ol4GkPlMzi8jBb4CrK6c1fp9cvzw+pLuLB/a+eVW++DViA/QmlOtU7bUekPS+sCPyCaMI8kskfuTQeFOYB9JjwOPkrmIXgaIiNMbUf72mk9z48fJXFPjImKKpFlkuuCWiHhC0onk1dzOi7phzM2u9KO9AWwo6bGImFEO5G+Rqctfo1yKtJwNbBERV0j6fXn9K9D4JtrK1vRjXl7uCZKWLQF7KDmBAnKEzmvADiXI30VOjHrX5eVUd3Wb3iy/Na/ygz9BmSCv1gl7BnBCRBwJEBFXk7ntP0Om7riIPAhcC1wcEZMbUfaOKM1VtSbRz0r6nKQVIuI35D7Vrk97Ljl7dkPlsMvfkwe+2X3lDKauGeYCYCzZQQtA5JW6biHP2rYuFcXfAruWQR6vRsQrdc2+DW0JqExHrqShUTdWXtI25IiJO8jhYbWc1MPIMeLTJH2NzPtRy+o3LCJe6u2yW9/Q+kctabOIuEE5jPeL5GzM2WQ6jbuB58jRYN+OiNtKp+6sRgeF+SkBeiTwrYj4alm2OPBT4ANkcr/3A/9D5oE/n5x09XdJB5Jt3cdExGMNKH6HaF6ai7byYR1NVv5+HjmhrLb8M8CnyMlX50bE+b1Z5vbq9zV9SeMk3QicJmmPsqx27cwJ5JWWtmZeNr/3Ad9RDr3ahLyy1ToR8XYZbdHv/2fWcXXt9rOV+VUAfiFpT+AxMsB/kZxcV5vLMZMM/muW4PJ6swb8mtJO/WlJE8uiMeSFbTaJiH3IGv5WZPPUFWTWVMhx6ac0e8BvfeBuNQqp1hz+S7LfZYO6544n+2EOjogtagFfTXhd3n5b0y896seQNfXDyQtAn0vmY3+ZvDjH4mSypzMjZz4OIGthB5OZMo8lv9TrR8Shvb0P1rcoL0L+Y+C2iDhWObv0f4DxUZcorrTvf5kcvjcw2pMDvYlIOoIcc/8pshPzDODQiHhI0jpkPqivkPmiriGvBNfjM027k6TtyaHYdwDXlSapucNRlTn+VyGbdHYCHiGHdU8v6zXt0Or+XGtdijwVnRQRl0TEr8mxwruReS8GksF/0xLwlyS/nFPJD3tHcnTOUWSua7O5WtfgSrv9r8j89scCRMS15IXpjyjrjJH0U3Jc/h9LX1CfCfiSPiHpbvIayx8k88W8Qo5uGw8QEX8ic1EtGRHTgI81c8Av/XqqezxK0rlk/p9vkc1UX1Rm0oV5g1/OI2v7E8j0CjtExPTatpo14EM/DvrlNPQaYHllLhOAZ8ip0K+Tp6FvAAcpLy5+A7CpMr3vHGBncjLM1hFxee/vgTWzunb7pcuiIWSb/QxJH5O0pzL970HAzspx2gPJdu9PRMRpjSh3e82nWWIL4DcRcQg5Q31PciLS74BtJf1Y0hVkoJwKEM199araIIyoa7oROTx7RkTcTPZXXE1WFolM6zwg8spuX46I8ZH57lXbXiP2pSP6TfOO3j3lu3YKNoystY8iaybrkadhb5JXoxlG1uhXAU6NiBvqtte0p2fWGPWdepImkGPtHyQrDH8nv0sbkWeGG5Lt+HuRF8V4MiK+18Zmm4rePSt9GHlxmzvI8fZnkXMKri2/rzPJdOHbK1NLbA68HRE/b1DxO0WZw39lcuTUVeTB7CsRsXZ5fktge3Ly2CttdOz2rVgRTZB3uis33p1venDr5WQ+8ivIdnvIXvefUIZgLmh7vvlGjkZZh7ySUa2SNJ5M/TyGzKcyg7zQff3376OUC5PXXteXbuQVqf5MDkU8m+zrOoxsv6/9H7Ymm0r/vdHlbec+tb4U4xLkAflEcsLVg8BXy/KbgO+W9b5MJk5r+D50x61PT85qVbv/KrCRpGMj4pHa8oj4o6SbyYtIfzgiHpV0MNlR+2TrbYXH2xtzmze+Qybb+yvZfHM72Uk5mMyuugPZDPiDiLhFmWrhA+Qs2s3JYNnUShv0uyYWKvNNQdZ4HwC+SQ5b/k/yrOYoSU+RZwH/S45OamqtYkWtZj6IPIM5mzwj+xdwc+SY+hOA30gaTY7oa/qztPbqc236kgZL+p2kUZEJ0YZKOgn4JHlkfqRu3VoHze/IGtvW5RR9djkwvFHX8eJgbwBI2gL4v/Lw42Rele8A/yVpUzLd7+fIisNWEXFCadsfVW6vkbMxT4aeycHUHeratOdIWlqZ1A3yzHgd8gwmyPH2K5D/i92BZ8n/ydWRKUie7v3Sd0zZx2ElVnyzDMkeTjbtXknm7N8oIv5U+mJ+R47EGhI5BPPmhhW+uzX6VKO9N5h3FXnyCzi83F+SPBV7f3k8cD6v3x5YvdH74Vvz38gmwTl1j4eWv4cAfyBHhl1NDrlchAyQfyRHcTS8/O3Yv/rLaw4kmzfuJMfSr1GW30TmsYesEX+OvAj7kEaXv537uFSrx2uRgfyr5BnMw+QlCU8hr2ddaw4+gsyYOxhYjRz8sXaj96c7b32iI1dSC3lVmXM17+LZ/wccEhEXlSFWPyfHR9c62gZFZvGbm77VrL0k/Qp4LSL21rtTAP8DmEiOUNmLDAzLAT+MiAsaVNx2qTurrf1GliBnpf8zIo6W9DvyLOUgcv7KBeTv7jllxsjaBcvfjCYNHGVo5eHk2ch04PqIuE7SeLLd/jpy7sRgsiK4PtlEV+u3mUFJi1y29zng9qhLktbXNXXQb9UO9zTZebYO+aP7BHk5uTWUOaunA5dGxFPK2YIREec2quzWt5XmmieBj0bEw5KGRcTrpYLxhyhDLiWtFhGPN7SwHSTpQ+RV344m+yvmkNkjRwL/JOcQnKS8VvHAiNitUWVtL+VlNk8hR+SdRl5adAtyxM3a5Giqb5ITxn4UEb8qr1ucPNBtACwSEbeX5f33AkWNPtWYz6mZaDWKhkxg9Dqwat2y+8hOs2WZd3WaW8ihV6s0ej9869s3sh3/zlbLfgOMbXTZurBPe5LDlnevW/YV4Lhy/0CyY3ZDshlrm0aXuZ37tQvvbpKrVWjPJM9OVgImAxPq1vk5bTTJMZ8m4v5ya8qO3EhzJK0g6Sfl1OxgslaySd2q+5O1lVmR1w49ikzotHVEPNHb5bb+JTIj5gqSNpW0fGn+eBl4sn4WZx/zMJkG+O91y0aQbd6QHdGPkbXelyIzZja9iLgYmFJG5kGOtoIcbvnp8vh04ABJF0l6gEx+d2Eb2+o7Y+47oWmad+qbcsrjvclOl0uBn0TEy2Vm7WVkgqdaG+vlwFsRsUur7fWtCRPWlCTtQgaGu8msin1q4lFbSnPoChGxa3m8Cjn0soU8e/56RMxoWAE7SZnS+FpyUMcslcy6pUnu3og4uYxQ+gTw94h4oLyu/zbltKHhQb+NzqXatWn/i8w3fjbZ1rhkRDwu6VKys+VachbdKcAy0cbFls26g6R9gfOiLjV3X1Y6Za8CjoyI6+uWLRMRjza0cF1U4sMrEfGFumXXAcdHq2GXbc1RqIKGB/0aSRuTlyGcQtasdiHbH2eQGTKXJZMcXUy2xS1BXnfzzvL6AVX78Mw6S9IBZKqBNRpdlu5U1wG/SkTMkHQk2T+xX5QMmFXXkBm5mneBgtrfieTwt6OAr5E5co4lO16mkz3uuwJLR8SrkvaPiJfrt+mAb9YhZwNzlOnEo780b0TETEn/Czwu6V5y6OYXHfDn6dWOXL33cmGLlb/DyItKLELOcnyMHHL1N7KzZV9yYsyz5fUv12/PzDomIt6MiDMiU4/0i4BfExFHkaP4/ici9oi8Cl5TDlpphF6t6ce87H3rkr3qQ8hp3auQw8YeBLaNiL8qry05mGzy2Qj4bET8ra3tmZnVi4jtavfd9PtuvdqmX2YAXkJOBplKXoBgezLof56cYTtVefWd/yFH70yPctWh/nYqamY9x8G+bT1W05/PkMkWMojvLWk4mbXvexGxjaRxwOmSZpIz6E6vH2vvD9DMOsLxom09FvTrmnJ2JpMW3UNO/KhdTPgtsjNpd0nbRcQRktYE1gD2j7wyTf32/AGamXVRtzTvtDW5QdIYMiXrS+SwyyHAF8nmnQsj4hfKK/OcR84I3CIi3qx7vSdXmZl1sy73aCsvMjC23K8/c1gTmBwRW5MpFJ4jL8RwGHCspF2BX5KpFZ4nL8hQ26Yc8M3Mul93DGPaghxPT2Qq41pOko+QaWchs9idQmbIfIjMmfNh4HJyqOabZE4QynbcUWtm1gO6HPQj4kzgr5IOLYsGl78/AdaVtHZEvE2OwX8amB0R15IXlZ5OXtjgeWBmH05iZWbWJ3RXm37rREdDIuLNciDYjUx7vCc52eor5YxgBfJqPHfUUimYmVnP6rZx+iXR0asRsXfdsoFkYF8N+Gv4oiZmZg3VnUG/lujogxHxQsl/sRZwYET8pW49j7c3M2uQbhunX5fo6ElJfyNn3H4hIqbCvGGdDvhmZo3T7WkYJF0B/LQuT7dr9mZmTaLHcu9U9QIFZmbNrEfSMNTV7j3e3sysiTTNlbPMzKzn+cICZmYV4qBvZlYhDvpmZhXioG/WQZKekrRsV9cxawQHfTOzCnHQt0qQ1CLpL5LOlvS4pPMl/bukOyT9VdKGkpaWdIWkByXdJWnt8tplJF0v6WFJPwdUt909JN0t6QFJp5d8U2ZNy0HfqmQV4EfAh8rtc+TFe74OHA4cA9wfEWuXx7UEgUcBt0fEGuQ1IFYCkPRhYBdgfESMBWYDu/fWzph1Ro9dI9esCT0ZEQ8BSHoYuCkiQtJDQAvwAWAHgIj4fanhLwFsAmxfll8j6cWyvQnA+sA95VIQi5JXiDNrWg76ViVv1t2fU/d4DvlbeLuD2xNwTkR8sxvKZtYr3LxjNs9tlOYZSZsCz0fEK8CtZFMQkrYCRpT1bwJ2lDSqPLe0pA/0cpnNOsQ1fbN5jgbOkvQg8C9gYll+DHBhaRK6E/gHQEQ8IukI4HpJA8gzhQPJtOJmTcm5d8zMKsTNO2ZmFeKgb2ZWIQ76ZmYV4qBvZlYhDvpmZhXioG9mViEO+mZmFeKgb2ZWIf8Pocb+3rwqop4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,20))\n",
    "df_models.plot.bar(x='model',rot=30,stacked=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "54aa72df",
   "metadata": {},
   "source": [
    "## Improving Model performence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "5082597d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.496377Z",
     "start_time": "2021-05-03T14:04:21.405621Z"
    }
   },
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "d6c5fcf9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.512334Z",
     "start_time": "2021-05-03T14:04:21.498377Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "osVerDimId_cr                   -0.164867\n",
       "browserId_cr                    -0.152837\n",
       "siteId_cr                       -0.125855\n",
       "pubclientid_cr                  -0.121393\n",
       "itemid_cr                       -0.084807\n",
       "position_cr                     -0.058909\n",
       "uuidSource_cr                   -0.045874\n",
       "conversionDurationInMillis_cr   -0.045379\n",
       "connTypeDimId_cr                -0.034342\n",
       "itemcolumbiaid_cr               -0.029064\n",
       "ispDimId_cr                     -0.026533\n",
       "cityId_cr                       -0.016385\n",
       "geodimid_cr                     -0.013087\n",
       "countryDimId_cr                 -0.009699\n",
       "modelDimId_cr                    0.000109\n",
       "adslotdimid_cr                   0.021424\n",
       "algo_cr                          0.021767\n",
       "goalid_cr                        0.037235\n",
       "pricingtype_cr                   0.041786\n",
       "osId_cr                          0.059065\n",
       "geoGrpDimId_cr                   0.059635\n",
       "stateId_cr                       0.073453\n",
       "clientid_cr                      0.074817\n",
       "adLogType_cr                     0.108480\n",
       "lookUpFrom_cr                    0.108480\n",
       "goalTypeId_cr                         NaN\n",
       "Name: conversion_fraud, dtype: float64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.corr()[\"conversion_fraud\"][:-1].sort_values()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ef24ebd",
   "metadata": {},
   "source": [
    "# After Cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "778b21bc",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.527295Z",
     "start_time": "2021-05-03T14:04:21.514335Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "osVerDimId_cr                   -0.164867\n",
       "browserId_cr                    -0.152837\n",
       "siteId_cr                       -0.125855\n",
       "pubclientid_cr                  -0.121393\n",
       "itemid_cr                       -0.084807\n",
       "position_cr                     -0.058909\n",
       "uuidSource_cr                   -0.045874\n",
       "conversionDurationInMillis_cr   -0.045379\n",
       "connTypeDimId_cr                -0.034342\n",
       "itemcolumbiaid_cr               -0.029064\n",
       "ispDimId_cr                     -0.026533\n",
       "cityId_cr                       -0.016385\n",
       "geodimid_cr                     -0.013087\n",
       "countryDimId_cr                 -0.009699\n",
       "modelDimId_cr                    0.000109\n",
       "adslotdimid_cr                   0.021424\n",
       "algo_cr                          0.021767\n",
       "goalid_cr                        0.037235\n",
       "pricingtype_cr                   0.041786\n",
       "osId_cr                          0.059065\n",
       "geoGrpDimId_cr                   0.059635\n",
       "stateId_cr                       0.073453\n",
       "clientid_cr                      0.074817\n",
       "adLogType_cr                     0.108480\n",
       "lookUpFrom_cr                    0.108480\n",
       "goalTypeId_cr                         NaN\n",
       "Name: conversion_fraud, dtype: float64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# plt.figure(figsize=(10,6))\n",
    "# df.corr()[\"conversion_fraud\"][:-1].sort_values().plot(kind='bar')\n",
    "df.corr()[\"conversion_fraud\"][:-1].sort_values()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d592782f",
   "metadata": {},
   "source": [
    "# Before cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "98171c3a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.822506Z",
     "start_time": "2021-05-03T14:04:21.528293Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,10))\n",
    "df.corr()[\"conversion_fraud\"][:-1].sort_values().plot(kind='bar')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "edd2ab83",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.852940Z",
     "start_time": "2021-05-03T14:04:21.825498Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "osVerDimId_cr                   -0.164867\n",
       "browserId_cr                    -0.152837\n",
       "siteId_cr                       -0.125855\n",
       "pubclientid_cr                  -0.121393\n",
       "itemid_cr                       -0.084807\n",
       "position_cr                     -0.058909\n",
       "uuidSource_cr                   -0.045874\n",
       "conversionDurationInMillis_cr   -0.045379\n",
       "connTypeDimId_cr                -0.034342\n",
       "itemcolumbiaid_cr               -0.029064\n",
       "ispDimId_cr                     -0.026533\n",
       "cityId_cr                       -0.016385\n",
       "geodimid_cr                     -0.013087\n",
       "countryDimId_cr                 -0.009699\n",
       "modelDimId_cr                    0.000109\n",
       "adslotdimid_cr                   0.021424\n",
       "algo_cr                          0.021767\n",
       "goalid_cr                        0.037235\n",
       "pricingtype_cr                   0.041786\n",
       "osId_cr                          0.059065\n",
       "geoGrpDimId_cr                   0.059635\n",
       "stateId_cr                       0.073453\n",
       "clientid_cr                      0.074817\n",
       "adLogType_cr                     0.108480\n",
       "lookUpFrom_cr                    0.108480\n",
       "goalTypeId_cr                         NaN\n",
       "Name: conversion_fraud, dtype: float64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.corr()[\"conversion_fraud\"][:-1].sort_values()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1da00196",
   "metadata": {},
   "source": [
    "## Test\n",
    "- As we observe, Random Forest perform best\n",
    "- so, we will prepare the test Data.\n",
    "- "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "2f300e19",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.900842Z",
     "start_time": "2021-05-03T14:04:21.855931Z"
    }
   },
   "outputs": [],
   "source": [
    "### TEST DATA\n",
    "df1 = pd.read_csv(\"Test Data.csv\")\n",
    "\n",
    "# Cleaning data\n",
    "# Colum del\n",
    "df1.drop([\"clmbuserid_cr\",\"conversiontime_cr\",\"imprid_cr\",\"audiences_cr\",\"clickbid_cr\",\"ip_cr\",\"refurl_cr\",\"templateid_cr\",\"templateid_cr\",\"time_cr\",\"v_cr\",\"allAudiences_cr\",\"impressionTimeInMillis_cr\",\"stateGrpDimId_cr\",\"deviceId_cr\",\"uvh_cr\",\"uv_cr\",\"platformId_cr\",\"sdkVersion_cr\",\"usrClusterId_cr\",\"cityGrpDimId_cr\",\"siteClusterIds_cr\",\"refClusterId_cr\",\"paid_cr\",\"spend_cr\",\"attributionType_cr\",\"conversionid_cr\",\"optimize_on_cr\",\"bundleId_cr\",\"clickTimeInMillis_cr\"],axis=1, inplace=True)\n",
    "\n",
    "## ,\"connTypeDimId_cr\"\n",
    "\n",
    "# Convert Values to INT in Y\n",
    "\n",
    "# Droping Null value\n",
    "df1 = df1.dropna(how='any',axis=0) \n",
    "\n",
    "\n",
    "# Dividing Dataset INTO X and Y\n",
    "X1 = df1.iloc[:,1:].values\n",
    "index = df1.iloc[:,0].values\n",
    "X1 = sc.transform(X1)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "d99e8909",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:21.916768Z",
     "start_time": "2021-05-03T14:04:21.901807Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>450</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>451</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>452</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>453</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>454</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>455 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     0\n",
       "0    0\n",
       "1    0\n",
       "2    0\n",
       "3    0\n",
       "4    0\n",
       "..  ..\n",
       "450  0\n",
       "451  0\n",
       "452  0\n",
       "453  0\n",
       "454  0\n",
       "\n",
       "[455 rows x 1 columns]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y1= LR_classifier.predict(X1)\n",
    "y1_df = pd.DataFrame(y1)\n",
    "y1_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "049680b2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.742654Z",
     "start_time": "2021-05-03T14:04:21.918762Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'y2_df' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-30-a4cfcd862da4>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0my2_df\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'y2_df' is not defined"
     ]
    }
   ],
   "source": [
    "y2_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a27c3696",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.753633Z",
     "start_time": "2021-05-03T14:03:57.839Z"
    }
   },
   "outputs": [],
   "source": [
    "index = df1[\"record_id\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97b0e264",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.754630Z",
     "start_time": "2021-05-03T14:03:57.841Z"
    }
   },
   "outputs": [],
   "source": [
    "horizontal_stack = pd.concat([index, y2_df], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f441f359",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.755627Z",
     "start_time": "2021-05-03T14:03:57.843Z"
    }
   },
   "outputs": [],
   "source": [
    "horizontal_stack"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "957a68ae",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.757622Z",
     "start_time": "2021-05-03T14:03:57.845Z"
    }
   },
   "outputs": [],
   "source": [
    "horizontal_stack['conversion_fraud'] = np.where(horizontal_stack[0]==0, 'FALSE', 'TRUE')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5977c4d2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.758619Z",
     "start_time": "2021-05-03T14:03:57.849Z"
    }
   },
   "outputs": [],
   "source": [
    "horizontal_stack"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a9627098",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.760615Z",
     "start_time": "2021-05-03T14:03:57.851Z"
    }
   },
   "outputs": [],
   "source": [
    "horizontal_stack.drop([0])\n",
    "horizontal_stack"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca445ab2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.761612Z",
     "start_time": "2021-05-03T14:03:57.853Z"
    }
   },
   "outputs": [],
   "source": [
    "horizontal_stack.to_csv(\"submissio.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2d6bf437",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.764603Z",
     "start_time": "2021-05-03T14:03:57.856Z"
    }
   },
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a3aa9777",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2021-05-03T14:04:22.765601Z",
     "start_time": "2021-05-03T14:03:57.858Z"
    }
   },
   "outputs": [],
   "source": [
    "filename = 'finalized_model.sav'\n",
    "pickle.dump(model, open(filename, 'wb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cbee160d",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "tf-gpu",
   "language": "python",
   "name": "tf-gpu"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "position": {
    "height": "462.4px",
    "left": "1087px",
    "right": "20px",
    "top": "107px",
    "width": "350px"
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}